AERODYNAMICS
[AERODINAMIKA]
Authorised by the Ministry of Higher and Secondary Specialized Education, USSR as a Textbook for the Students of Polytechnics
N.F. Krasnov
Published for the National Aeronautics and Space Administration and the National Science Foundation, Washington, D.C. by Amerind Publishing Co. Pvt. Ltd., New Delhi
1978
Translatedandpublishedfor the National Aeronautics and Space Administration, pursuant to an agreement with the National Science Foundation, Washington, D.C. by Amerind Publishing Co. Pvt. Ltd., 66 Janpath, New Delhi
Available from the U.S. Department of Commerce National Technical Information Service Springfield, Virginia 22161
UDC 533.6
In this book the basic concepts of modern aerodynamics are discussed. The question of the kinematics of a fluid medium and the basic equations of hydrodynamics for a complicated, chemically reactive, viscous, compressible model of fluid flow are discussed in detail. These questions are applicable to the problems of hypersonic flight, wing theories in subsonic and supersonic flows, supersonic flow past rotating bodies, aerodynamic interference, determination of frictional force, heat transfer and heatresistant sheaths. The basic concepts of the aerodynamics of dissociating gases are also treated. The book is recommended as a textbook for students in higher technical institutions and will also be useful to engineers and scientists.
PREFACE
Aerodynamics forms the theoretical basis of spacerocket and artillery technology and the foundation of aerodynamic calculations for modern flight vehicles. The basic concepts of aerodynamics are used in the investigation of an outer flow over different bodies and the motion of air (gas) past some of these structures. Therefore it is difficult to be a good engineer in the field of aviation, artillery, rocketry, automobile transport, internal combustion engines, etc. without a thorough knowledge of aerodynamics. To be a specialist one must acquire knowledge of aerodynamics where air or gas flow phenomena occur in some form. In this book the applications of aerodynamics particularly in rocket technology and modern highspeed aviation are studied along with the general laws of motion of a gas medium. In following any course embracing aerodynamics a deep understanding of its theoretical principles is very important. Without it the actual solution of practical problems, scientific investigations and research work become impossible. Therefore special attention must be devoted to mastering the material in the first six chapters of this book. These cover the following: basic concepts and definitions of aerodynamics, kinematics of a fluid medium, equations of motion of a gas in the general case, where variations in its physicochemical properties are considered, the theory of shock waves, the method of characteristics, which is very widely used in the investigation of supersonic flows, and finally the general theory of the motion of a gas in twodimensional space (known as the theory of twodimensional motion). The scientific information given in the remaining chapters of this book shows the practical applications or connections with the formulation of methods of aerodynamic calculations for flight vehicles and their individual elements. This is also useful if we are to achieve a deep understanding of the theoretical foundations of aerodynamics. This structuring of a course corresponds to the important principle of presenting scientific information according to a deep, thorough approach for the use of students as an active means of obtaining solutions of some practical problems. This method of study is based on the multipurpose approach of managing scientific information for easy retrieval and crossreference to the logical connections obtaining among different items of knowledge. It is worth noting here that the presentation of the methods of aerodynamic
calculations has a significant value in itself. It enriches the students' knowledge and steers them along the road of establishing the interrelation between the theoretical and practical solutions of specific problems besides acquainting them with the new phenomena related to the processes of the interactions of bodies. Here the aerodynamics of isolated wings (lifting surfaces) and rotating bodies (cones) are first considered and then the aerodynamic calculations of flight vehicles in the form of various combinations of wings, control surfaces and rotating bodies including mutual interference are taken up. This layout is quite normal for a textbook aimed at a complete and uptodate understanding of the problems of applied aerodynamics. In this book particular attention is given to the highlights of important problems of theoreticaI and applied modern aerodynamics of high speeds associated with the investigation of friction, heat transfer, mass ablation, force and heat interactions under the condition of motion of a body in a dissociating medium. The thermodynamic and kinematic parameters of a dissociating gas and the basic physical properties of a highly dissociating medium are also discussed as applied to the above problems. Obviously not all the problems pertaining to aerodynamics can be covered in a textbook. The scientific information presented in this book is what is necessary for any specialist engaged in scientific and engineering activities in the field of aviation and spacerocket technology. The content and volume of this information will be enough for the understanding of other problems of aerodynamics faced by young specialists in practice. Among the problems not treated in this book are, in particular, magnetoaerodynamic investigations, determination of an unsteady flow over bodies, the problems of experimental aerodynamics and the optimum shapes of flight vehicles. The author will be very satisfied if the mastery of the material presented in thisbook serves as an aid for individual work and detailed study of modern aerodynamics. The book is written on the basis of teaching experience in aerodynamics following the teaching program in the N.E. Bauman Moscow Higher Technical Institute. It is recommended for students of higher educational institutes and of faculties for special training in the field of flight vehicles. It may also be useful to workers in the affiliated scientific research institutes, design bureaus and production plants. A special conversion table is provided at the end of this book for convenience in converting the units of measurement of physical quantities used here to the corresponding new units accepted in the International System (SI). The leading scientists and technologists of the RSFSR, Prof. N.S. Arzanikov and Prof. I.P. Ginzburg, and colleagues of their section carefully went through the manuscript of this book. They made many useful comments and came up with valuable suggestions for the improvement of the contents and presentation of the manuscript, which were highly appreciated during
preparation of this book. The author acknowledges his deep indebtedness to all of them. The author appreciates that this book is not necessarily free from defects and will gratefully accept all comments and suggestions from readers.
The following list gives the basic nota?ions popularly used in the field of aerodynamics in the Soviet Union and the United States of America. This will help readers who are used to American and English notations to understand the material in this book. The typical coordinate systems used in these countries are reproduced in Figs. 1 and 2.
In the USA and England In the
USSR
tR
aspect ratio sound velocity critical sound velocity wing span cross sectional area lift coefficient side wind force coefficient drag coefficient pitching moment coefficient rolling moment coefficient yawing moment coefficient specific heat at constant pressure specific heat at constant volume chord acceleration due to gravity altitude enthalpy rolling moment pitching moment yawing moment mass load factor angular velocity about X axis angular velocity about Y axis (Z axis in the USSR) angular velocity about Z axis, (Y axis in the USSR)
wing planform area entropy semispan absolute temperature thrust of propeller or engine maximum thickness of wing velocity component of velocity along X axis component of velocity along Y axis (Z axis in the USSR) component of velocity along Z axis ( Y axis in the USSR) weight of airplane lift drag crosswing force angle of attack angle of side slip ratio of specific heats circulation boundary layer thickness relative density downwash angle angle of elevator deflection angle of pitch sweepback angle coefficient of viscosity kinematic coefficient of viscosity mass density time ratio stream function velocity potential
CONTENTS
Preface
. . . . . . . . . . . . . . . . . . . . .
. . . . . . .
ix
Introduction
1. Forceinteraction of the Medium on a Moving Body . . 1.1 Surface force . . . . . . . . . . . . . . 1.2 Properties of pressure in an ideal fluid . . . . . . 1.3 Effect of viscosity on fluid motion . . . . . . . 2. Resultant Forceinteraction . . . . . . . . . . . 3 . Determination of Aerodynamic Forces and Moments from a Given Distribution of Normal Pressure and Shear Stress. Concept of Aerodynamic Coefficients . . . . . . . 4. Basic Properties of Highspeed Gas Flow . . . . . . 4.1 Compressibility of gas . . . . . . . . . . . . 4.2 Heating of gas . . . . . . . . . . . . . . 4.3 Condition of air at high temperatures . . . . . . 5. Basic Relations for a Diatomic Dissociating Gas . . . . 5.1 Degree of dissociation . . . . . . . . . . . 5.2 Equation of state . . . . . . . . . . . . . 5.3 Thermodynamic relations . . . . . . . . . . 5.4 Coefficient of dynamic viscosity . . . . . . . . 5.5 Mixture of diatomic gases . . . . . . . . . .
Chapter 2 Kinematics of a Fluid Medium . . . . . 1. Methods of Kinematic Investigatioil of a Fluid 1.1 Lagrangian method . . . . . . . . 1.2 Eulerian method . . . . . . . . . 1.3 Stream line and streak line . . . . . . 2. Analysis of Motion of Fluid Particles . . . 3. Irrotational Motion of a Fluid . . . . . . 4 . Equation of Continuity . . . . . . . . 4.1 General form of equation of continuity . 4.2 Cartesian system of coordinates . . . . 4.3 System of curvilinear coordinates . . .
. . . . . . . . . . . .
. . . . .
. . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
4.4 Equation of continuity for gas flow over curvilinear surface . . . . . . . . . . . . . . . . 4.5 Equation of mass flow . . . . . . . . . . 5. Stream Function . . . . . . . . . . . . . 6. Vortex Lines . . . . . . . . . . . . . . . 7. Circulation of Velocity . . . . . . . . . . 7.1 Concept of the circulation of velocity . . . . . 7.2 Stokes' theorem . . . . . . . . . . . . . 7.3 Velocities induced by vortices . . . . . . . . 8. Complex Potential . . . . . . . . . . . . . 9. Typical Fluid Flows . . . . . . . . . . . . . 9.1 Uniform flow . . . . . . . . . . . . . . 9.2 Plane source and sink . . . . . . . . . . . 9.3 Threedimensional source (sink) . . . . . . . 9.4 Dipole . . . . . . . . . . . . . . . 9.5 Flow with circulation (vortex) . . . . . . .
. . . .
. . . .
Chapter 3 Foundations of Fluid and Gas Dynamics . . . . . 1. Equations of Motion of a Viscous Fluid . . . . . . 1.1 Cartesian coordinates . . . . . . . . . . 1.2 Equation of motion in vector form . . . . . . 1.3 Curvilinear coordinates . . . . . . . . . . 2. Equation of Energy and Diffusion of Gas . . . . . . 2.1 Equation of diffusion . . . . . . . . . . . . 2.2 Equation of energy . . . . . . . . . . . . 3. System of Gas Dynamic Equations. Initial and Boundary Conditions . . . . . . . . . . . . . . . . . 4. Integrals of Equations of Motion in an Ideal Fluid . . . 5. Aerodynamic Similarity . . . . . . . . . . . . . 5.1 Concept of similarity . . . . . . . . . . . . 5.2 Similarity criteria to account for viscosity and heat conductivity effects . . . . . . . . . . . . . 6. Isentropic Gas Flows . . . . . . . . . . . . . 6.1 Form of gas jet . . . . . . . . . . . . . . 6.2 Flow velocity . . . . . . . . . . 6.3 Pressure. density and temperature . . . . . . . 6.4 Gas flow from reservoir . . . . . . . . . . . 6.5 Motion of incompressible gas (fluid) . . . . . . . 6.6 Dissociating gas flow from reservoir (onedimensional problem) . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. .
. . . . . . . . . . . . . . .
. .
.
.
. .
.
. . . . . . . . .
CONTENTS XV
2.1 Oblique shock . . . . . . . . . . . . . . . 2.2 Normal shock . . . . . . . . . . . . . . . . 3. Oblique Shock in a Gas Flow with Constant Specific Heats . 3.1 System of equations . . . . . . . . . . 3.2 Formulas for calculation of gas parameters behind shock . 3.3 Angle of inclination of oblique shock . . . . . . . 4. Hodograph of Velocity . . . . . . . . . . . . . . 5. Normal Shock in Gas Flow with Constant Specific Heats . . 6. Shock at Very High Supersonic Spegds and Constant Specific Heats of Gas . . . . . . . . . . . . . . . . . 7. Solution of Problem of Shock in a Gas Flow with Variable Specific Heats taking into Account Dissociation and Ionization . 8. Shock Wave in Pure Dissociating Diatomic Gas . . . . . 9. Relaxation Phenomena . . . . . . . . . 9.1 Concept of nonequilibrium flows . . . . . . . . . 9.2 Equation of rate of chemical reactions . . . . . . . 9.3 Time of relaxation . . . . . . . . . . . . . . 9.4 Equilibrium processes . . . . . . . . . . . . . 9.5 Effects of relaxation in shock waves . . . . . . . . Chapter 5 Method of Characteristics . . . . . . . . . . . . 1. Equations for Velocity Potential and Stream Function . . . 2. Cauchy's Problem . . . . . . . . . . . . . . 3. Characteristics . . . . . . . . . . . . . . . . . 3.1 Conditions of existence . . . . . . . . . . . . 3.2 Existence of characteristics . . . . . . . . . . . 3.3 Property of orthogonal conjugate characteristics . . . 3.4 Transformation of equations of characteristics in hodograph plane of velocity . . . . . . . . . 3.5 Equations of characteristics in the hodograph plane for particular cases of gas motion . . . . . . . . . . 4 . Scheme of Solving Gas Dynamic Problems by Method of Characteristics . . . . . . . . . . . . . . 5. Application of Method of Characteristics to Solution of the Problem of Contour Design of Nozzle of a Supersonic Aerodynamic Wind Tunnel . . . . . . . . . . . . . .
Chapter 6 An Airfoil and a Finite Wing in an Incompressible Fluid Flow 1. Thin Airfoil in Incompressible Flow . . . . . . . . . 2. Cross Flow over Thin Plate . . . . . . . . . . . . 3. Thin Plate at an Angle of Attack . . . . . . . . . . 4. Finite Wing in Incompressible Fluid . . . . . . . . . Chapter 7 Wing Profile fn a CompressibIe Gas Flow . . . . . . . 1. Subsonic Compressible Flow over Thin Airfoil . . . . . .
xvi
CONTENTS
&
1.1 Linearization of equation of velocity potential . . . . 1.2 Relationship between compressible gas flow and incompressible fluid flow over thin airfoil . . . . . . . . . 2. Academician S.A. Khristianovich's Method . . . . . . . 2.1 Description of method . . . . . . . . . . . . . 2.2 Conversion of pressure coefficient from incompressible fluid flow to Mach number iVL >O . . . . . . . . 2.3 Conversion of pressure coefficient from Mach number M, I >O to Mach number M, 2 >0 (M, I #M, 2) . . . . 2.4 Determination of critical Mach number M . . . . 3. Wing Profile in Flow having SupercriticalVelocity (M, >MWor) 4. Flat Plate in Supersonic Flow of Gas having Constant Heat Coefficients . . . . . . . . . . . . . . . . . . 5. Supersonic Flow Past Airfoil of Arbitrary Form . . . . . 5.1 Application of method of characteristics . . . . . . 5.2 Hypersonicflow over thin profile . . . . . . . . . 5.3 Thin profile in small perturbation flow . . . . . . . 5.4 Aerodynamic coefficients . . . . . . . . . . . 6. Swept Wing of Infinite Span . . . . . . . . . . . . .7. Airfoil in Supersonic Gas Flow with Variable Heat Coefficients
Chapter 8 A Wing in a Supersonic Flow . . . . . . . . . . I . Linearized Theory of Finite Wing in Supersonic Flow . . . 1.1 Linearization of equation for a potential function . . . 1.2 Boundary conditions . . . . . . . . . . . . 1.3 Components of resultant velocity potential and aerodynamic coefficients . . . . . . . . . . . . . 1.4 Characteristicsof supersonic flowinteraction of wings . 2 Method of Sources . . . . . . . . . . . . . . . 3. Delta Wing with Symmetrical Profile (a =0, c, =0) . . . . 3.1 Semiwing with subsonic leading edge . . . . . . . 3.2 Delta wing symmetrical about x axis with subsonic leading edge . . . . . . . . . . . . . . . . . . . 3.3 Infinite semiwing with supersonic edge . . . . . . . 3.4 Delta wing symmetrical about x axis with supersonic leading edge . . . . . . . . . . . . . . . . 4. Flowinteraction of Swept Wing with Four Corners having Symmetrical Profile and Subsonic Edges at Zero Angle of Attack 5. Flowinteraction of Wing with Four Corners. Symmetrical Airfoil and Edges of Different Types (Subsonic and Supersonic) 5.1 Leading edge and middle edge subsonic. trailing edge supersonic . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . .
.
5.2 Leading edge subsonic. middle edge and trailing edge supersonic . . . . . . . . . . . . . . . . 5.3 A 1 edges of wing supersonic . . . . . . . . . 1 5.4 Delta wing . . . . . . . . . . . . . . . 6. Zone of Application of Method of Sources . . . . . . 7 Method of Dipoles . . . . . . . . . . . . . . 8. Flowinteraction of Delta Wing with Subsonic Leading Edge 9. Wing with Six Corners having Subsonic Leading Edges and Supersonic Trziiling Edges . . . . . . . . . . . . 10. Wing with Six Corners having Supersonic Leading and Trailing Edges . . . . . . . . . . . . . . . . 11. Drag of Wings with Subsonic Leading Edges . . . . . 11.1 Suction force . . . . . . . . . . . . . . 11.2 Induced drag . . . . . . . . . . . . . . 12. Aerodynamic Characteristics of Wing of Rectangular Plan . 13. Method of Reversal . . . . . . . . . . . . . .
Chapter 9 Cone in a Supersonic Flow . . . . . . . . . . 1. System of Equations of Axisymmetrical Flowinteraction of SharpNosed Cone . . . . . . . . . . . . . . 2. Flowinteraction of Cone at Constant Specific Heats . . . 3. Effect of Balanced Dissociation and Ionization of Gas on Flowinteraction of Cone . . . . . . . . . . . . 4. Blunt Cone . . . . . . . . . . . . . . . . . 4.1 Form of blunt noses . . . . . . . . . . 4.2 Special features of supersonic flow . . . . . . 4.3 Flowinteraction of cone with spherical blunt nose . . 5. Special Cases of Flowinteraction . . . . . . . . . 5.1 Unbalanced gas flows . . . . . . . . . . . . 5.2 Flows with radiation . . . . . . . . . . Chapter 10 SharpNosed Body of Revolution in a Supersonic Flow . . 1. Application of Method of Characteristics . . . . . . . . 2. Linearization of Equations of Flowinteraction for Thin Bodies of Revolution . . . . . . . . . . . . . . . . . 3. Axisymmetrical Flowinteraction . . . . . . . . . . 4. Nonaxisymmetrical Flowinteraction . . . . . . . . . 5. Calculation of Aerodynamic Coefficients . . . . . . . . Chapter 11 Aerodynamic Interference . . . . . . . I . Nature of Aerodynamic Interference . . . . . 1.1 Interference of wingbody combination . . . 1.2 Interference between wing and tail . . . . 2. Lift for Combination of Fuselage and Plane Wing
425 425 438 444 456 461 471 471 471 474 476
. . . . . . . . . . . .
. . . .
. . . .
xviii CONTENTS
2.1 Concept of coefficients of interference . . . . . . . 2.2 Determination of velocity potential . . . . . . . . 2.3 Velocity and pressure on fuselage with wing . . . . . 2.4 Velocity and pressure on wing in presence of fuselage . . 2.5 Determination of interference coefficients . . . . . . 2.6 Center of pressure . . . . . . . . . . . . . . 2.7 Lift force of wingfuselage combination . . . . . . . 3. Effect of Angle of Banking on Interference between Fuselage and Plane Wing . . . . . . . . . . . . . 3.1 General relation for coefficient of pressure . . . . . . 3.2 Pressure over fuselage . . . . . . . . . . . . . 3.3 Pressure over wing . . . . . . . . . . . . . . 3.4 Lift force and center of pressure . . . . . . . . . 3.5 General relations of forces and moments for plain combination with banking . . . . . . . . . . . . . 4. Cruciform Combination . . . . . . . . . . . . . 4.1 Pressure and lift force . . . . . . . . . . . . . 4.2 Coefficients of interference and center of pressure . . . 4.3 General relations for forces and moments . . . . . . 4.4 Calculation of interference of flight vehicles consisting of "thick" bodies and wings . . . . . . . . . . . . 5. Interference between Wing and Tail . . . . . . . . . 5.1 General definitions . . . . . . . . . . . . . . 5.2 Subsonic speeds . . . . . . . . . . . . . . . 5.3 Supersonic speeds . . . . . . . . . . . . . . 5.4 Co&cient of interference . . . . . . . . . . . 6. Control Elements . . . . . . . . . . . . . . . . 6.1 Basic types of control elements . . . : . . . . . 6.2 Aerodynamic analysis of controls . . . . . . . . .
Chapter 12 Friction . . . . . . . . . . . . . . . . . 1. Boundary Layer Equations . . . . . . . . . . . . 2. Generalized Equation of Boundary Layer . . . . . . . 2.1 Generalized equation of boundary layer in differential form 2.2 Integral relation of boundary layer . . . . . . . . 2.3 Characteristic thicknesses of boundary layer . . . . . 3. Laminar Boundary Layer over Flat Plate . . . . . . . . 4. Turbulent Boundary Layer over Flat Plate . . . . . . . 4.1 Application of logarithmic velocity distribution . . . . 4.2 Index law for velocity distribution . . . . . . . . . 5. Temperature and Enthalpy in Boundary Layer in Presence of Heat Transfer . . . . . . . . . . . . . . . . . 5.1 Distribution oftemperature and enthalpy . . . : . .
CONTENTS
xix
5.2 Characteristic temperature . . . . . . . . . . . 6 Application of Characteristic Parameters for Boundary Layer Calculations over Flat Plate at High Speeds of Flowinteraction 6.1 Laminar boundary layer . . . . . . . . . . . 6.2 Turbulent boundary layer . . . . . . . . . . . 6.3 Friction on cone at supersonic speeds of flowinteraction . 7. Effect of Axial Pressure Gradient on Friction . . . . . . 8. Mixed Boundary Layer. Critical Reynolds Number . . . . Chapter 13 Heat Transfer . . . . . . . . . . . . . . . 1. Aerodynamic Heating . . . . . . . . . . . . . . 1.1 Equation of heat balance . . . . . . . . . . . 1.2 Heat inflow from heated gas . . . . . . . . . . 1.3 Solar and earth radiation . . . . . . . . . . . 1.4 Radial heat flow from wall surface . . . . . . . . 2. Relation between Friction and Heat Transfer . . . . . . 3 Heat Transfer in Laminar Boundary Layer over Curved Surface 3.1 Arbitrary form of surface . . . . . . . . . . . . 3.2 Hemisphere . . . . . . . . . . . . . . . . 3.3 Blunt cone . . . . . . . . . . . . . . . . . 4 Heat Transfer by Diffusion . . . . . . . . . . . . 5 Wall Temperature . . . . . . . . . . . . . . . 5.1 Radiational equilibrium temperature . . . . . . . . 5.2 Equilibrium temperature in presence of additional sources of input and output of heat . . . . . . . . . . . . Chapter 14 Thermal Protection for Flight Vehicles . . . . . . . 1 Methods of Thermal Protection . . . . . . . . . . . 2. Determination of Heat of Ablation . . . . . . . . . 2.1 Sublimation, complete evaporation or carbonization (pyrolysis) . . . . . . . . . . . . . . . . . 2.2 Simultaneous melting and sublimation (evaporation) . . 2.3 Melting . . . . . . . . . . . . . . . . . 3. Mass Transfer and Effectiveness of Thermal Protection Shield . 4. Example for Calculation of Thickness of HeatProtection Shield Chapter 15 Aerodynamics of Rarefied Medium . . . . . . . . 1. Limits of Applicability of Theory of Motion of a Continuous Medium . . . . . . . . . . . . . . . . . . . 1.1 Molecular free path length . . . . . . . . . . . 1.2 Regimes of gas flow . . . . . . . . . . . . . 2. Pressure and Friction in Free Molecular Flow . . . . . 2.1 Scheme of interaction of molecules with wall . . . . . 2.2 Mass transfer . . . . . . . . . . . . . . . 2.3 Pressure . . . . . . . . . . . . . . . . .
XX
CONTENTS
2.4 Frictional stress . . . . . . . . . . . . 2.5 Transfer of kinetic energy . . . . . . . . 3. Adaptability . . . . . . . . . . . . . . 3.1 Exchange of momentum . . . . . . . . . 3.2 Exchange of energy . . . . . . . . . . . 4. Aerodynamic Forces . . . . . . . . . . . . 4.1 General expression for drag forces . . . . . 4.2 Cone . . . . . . . . . . . . . . . 4.3 Cylinder . . . . . . . . . . . . . . 4.4 Flat plate . . . . . . . . . . . . . . 5. Heat Transfer . . . . . . . . . . . . . . 5.1 Temperature of reflected molecules . . . . . 5.2 Calculation of heat transfer and wall temperature
Conversion Table for Units of Measurement Used in Aerodynamics from MKS System to International System (SI). GOST 986761 . . . 735 Bibliography
. . . . . . . . . . . . . . . . . . . .
737
INTRODUCTION
happens to be a part of mechanics, which is a science about the motion of bodies in general. It studies the laws of motion of air in relation to applied forces and on the basis of these forces establishes particular laws of interaction between the air and the body. The practical problems confronting man in connection with the flight of machines heavier than air served as a stimulus to the development of aerodynamics as a science. These problems were connected with the determination of forces and moments (known as aerodynamic forces and moments) acting on moving bodies. During this process the calculation of socalled supporting lift force was the main concern in the investigation of forceinteraction. In the early days of the history of aerodynamics it dealt only with low wind velocities as flight vehicles had low flight speeds. It is obvious that hydrodynamicsthe science of the motion of liquid (incompressible) particleswas the theoretical basis of aerodynamics. The foundation of this science was laid in the 18th century by L. Euler (17071783) and D. Bernoulli (17001783), members of the Academy of Science of Russia. In the scientific conference on "General Principles of Fluid Motion" (1755) L. Euler for the first time produced basic differential equations of motion of socalled ideal (inviscid) fluids and gases. Investigation of the fundamental law of hydrodynamics which establishes the relation between the pressure and velocity in an incompressible fluid goes to the credit of D. Bernoulli. He published this law in 1738 in his work Hydrodynamics. At relatively low flight speeds the effect of important properties of air, like compressibility, is infinitesimally small. However, the development of artillery and highspeed airplanes raised the problem of study of the laws of motion of air or, in general, of gas, at high speeds. It is found that if the forces acting on a body moving at high speed are calculated on the basis of the laws of motion of air at low speeds, these forces may be considerably different from the real values. The explanation of such phenomena was supplied by the investigation of the motion of air (gas) at high speeds resulting in changes in density with pressure. This change in density can often be
1
2 AERODYNAMICS
very appreciable. Along with this change in density there appears the property of compressibility of air. The compressibility of a gas is related to its internal energy which is to be taken into account when calculating the flow parameters determining the motion of the medium. The change in internal energy of the gaseous state and the total derivative of work done by the compressible gas during expansion are determined by the first law of thermodynamics. Thus thermodynamic relations must be applied in the aerodynamics of a compressible medium. If the gas medium moves at low speeds the heat content will be higher than the kinetic energy. In this case it is practically speaking unnecessary to calculate the change in heat energy with a change in speed (i.e. with a change in the kinetic energy of a fluid). Therefore it is not necessary to apply thermodynamic concepts and thermodynamic relations in the aerodynamics (hydrodynamics) of low speeds. Very high flight speeds, sometimes known as hypersonic speeds, are represented by the motion of rockets and space vehicles entering the dense layer of the atmosphere. At such speeds the surrounding gas or air not only undergoes a change in density but also a rise in temperature leading to various physicochemical changes in it. A considerable part of the kinetic energy due to the flight speed is transformed into heat and chemical energy. All these peculiarities in the motion of a gas medium resulted in the development of highspeed aerodynamics or gas dynamics as a special part of aerodynamics. Here we will be studying the laws of air (gas) motion at high subsonic and supersonic speeds and also the laws of interaction between the gas medium and the body moving through it at these speeds. Academician S.A. Chaplygin (18691942), one of the founders of gas dynamics, published the seminal work About Gas Jets in 1902. In this work he presented equations which form the theoretical basis of modern gas dynamics. These equations are known in science as "Chaplygin's equations." With the development of theoretical aerodynamics experimental aerodynamics came into existence. T deals with the experimental investigation of t the flow interaction between a body and a gas flowing around it with the help of various technical aids like wind tunnels and other equipment simulating the flow around an aircraft. The very first aerodynamic laboratories were built under the guidance of N.E. Zhukovskii (18471921) in Russia (at Moscow State University, the Moscow Higher Technical Institute and at Kuchin in the outskirts of Moscow). The N.E. Zlzukovskii Central Aerohydrodynamic Institute (TaAGI) was organized in 1918 with the help of V.I. Lenin on the initiative of N.E. Zhukovskii. This institute is now one of the most important centers in the world in the field of aerodynamics and has been named after N.E. Zhukovskii. With the development of aviation, artillery and rocket tkchniques the
INTRODUCTION
perfection of the theoretical basis of aerodynamics altered the nature of aerodynamic experimental techniques. The early comparatively small, lowspeed wind tunnels were superseded and the gigantic highspeed wind tunnel of TsAGI (1940), the modern hypersonic wind tunnels and the special tunnels with an artificially created supersonic flow of dissociating gas (namely wind tunnels with heated air and shock tubes, the plasmatunnel and others) came into being. The nature of the interaction between the gas medium and the body moving through it can be different for different conditions. The interaction at low speeds of motion is by nature basically of forces. With an increase in speed this forceinteraction is associated with surface heating due to the heat flow from the gas to the body, giving rise to heat interaction. At very high speeds the aerodynamic heating becomes so intense that it may lead to the destruction of the material of the walls of flight vehicles as a result of melting or sublimation. Consequently an ablation of the material and a change in the nature of the wall heating take place. Aerodynamic heating also leads to a chemical reaction between the solid wall and the gaseous medium around it as a result of which there appears the same effect, ablation of the body. High flight speeds can be the cause of ablation of the mass. The erosion of the material of the wall and damage to its structures can occur as a consequence of the mechanical interaction of the gas medium and the moving body. Investigation of all types of interactions between the gas medium and flight vehicles allows us to carry out aerodynamic analysis in relation to the quantitative estimations of criteria for a given interaction, in particular, to determine aerodynamic forces and moments, heat transfer and mass erosion (known as mass ablation). The given problem in modern conditions leads not only to the determination of total aerodynamic quantities such as resultant lift and drag forces, heat flow from dissociated gas to the surface, etc. but also to the estimation of distribution of aerodynamic quantities (both forces and heat) over the surface of the flying vehicle (e.g. normal pressure, frictional stress, local heat flows, local mass ablation). The solution of such problems requires a deeper investigation of gas motion than is necessary just to determine total aerodynamic interactions. This investigation includes determination of the gas parameters that characterize the motion at each point of the space it occupies and at every moment of time. Modern methods of investigation of the motion of a gaseous medium are based on a number of principles and hypotheses established in aerodynamics. One of them is the hypothesis of continuity of the moving gaseous medium, according to which the intermolecular spaces and molecular motions may be neglected and continuous variation of the basic gas parameters in space and time may be assumed. This hypothesis is derived from the condition that the free path lengths of molecules and amplitudes of their vibrational motion are
4 AERQDYNAMICS
negligibly small in comparison to the linear characteristic dimensions of flow, for example, wing span, diameter or length of cone, etc. The introduction of the hypothesis of continuity (solidity) should not contradict the concept of a gas medium even though it would seem that the medium must be incompressible in the absence of free molecular spaces. The postulation of a compressible continuous medium is based on the fact that the existence of molecular free spaces need not be taken into account in many investigations but the possibility of different degrees of concentration (density) may be assumed as a consequence of changes in these free spaces. In aerodynamic investigations the interaction between the gas medium and the body moving through it is determined on the basis of the principle of reversed motion. Thereby the interacting system of "stationary gas (air) medium and moving body" is replaced by the system of "moving gas medium and stationary body." In replacing one system with the other we have to observe the condition of the equality of the speed of the oncoming gas flow visavis the stationary body and the speed of this body in a stationary medium. This principle of reversed motion follows from the basic law of mechanics according to which the forces of interaction are independent of the circumstance which of the two interacting bodies (in this case gas and flight vehicle) is stationary and which in rectilinear uniform motion (law of relativity). From the practical point of view the principle of superposition is very helpful in calculation of the resultant aerodynamic characteristics (pressure, forces, heat flows, etc.). By the principle of superposition these characteristics can be resolved into their individual components. The calculation of each of the components requires the solution of an independent, simpler problem than the determination of the resultant values. In this way the total aerodynamic characteristics will take the form of superposed quantities, i.e. they will be obtained by the addition of respective components. The system of differential equations that forms the basis of the solution of flowinteraction problems in modern aerodynamics is treated separately as two main types of motions: one is free (external) inviscid flow and the other, flow in thin layers of gas, i.e. boundary layer flow where fluid motion is studied along with frictional effects. These divisions of flow are based on the hypothesis of the absence of adverse effects of the boundary layer on free flow. Then the parameters of the inviscid flow (i.e. at the external limit of the boundary layer) will be the same as those at the wall in the absence of this layer. Physical similarity conditions of flow interaction problems should be kept in mind when conducting aerodynamic experiments and calculations. The aerodynamic calculations of flying objects (rockets, airplanes) are generally based on preliminary investigations (theoretical and experimental) of flow over models.
INTRODUCTION
Aerodynamic similarity conditions must be observed in such investigations on models. These conditions are based on the characteristic, simple parameters determining the basic regimes of processes of investigation known as similarity parameters or criteria. The modern similarity problems and dimensional analysis widely used in aerodynamics are formulated in the fundamental work of Academician L.I. Sedov on Methods of Similarity and Dimensional Analysis in Mechanics. Aerodynamics is a manyfaceted science with a typical structure. The basic scientific directions and sections are more or less distinctly established in line with the fast developing aviation and space rocket technology. They are connected with the aerodynamic investigation of flight vehicles, individual structural elements and specific types of gas flows and the processes associated with them. It is natural that the whole classification of aerodynamics will to some extent be conditional because all these directions and sections or subsections are interrelated. But still, this "branchwise" specialization in aerodynamic science is of practical relevance. Let us study some typical directions and sections of modern aerodynamics. Two basic directions in which modern aerodynamics is developing can be distinguished. The first represents the socalled aerodynamics of forces associated with the solution of problems that deal with the forceinteraction of a medium. In other words this section of aerodynamics deals with the distribution of normal pressure and frictional stresses over the surface of flight apparatus and also the calculation of resultant forces and moments. The data so obtained are used in strength calculations for the structure of the bodies and individual elements and also in determination of its flight characteristics. The second direction includes the problems of science covering aerodynamics, thermodynamics and heat transfer and the flowinteraction with the heat effects being investigated. The heat flows from a heated gas to the wall are obtained and the wall temperature is determined. These data are necessary in strength calculations and in designing cooling systems for flight vehicles. In addition, the consideration of changes in the properties of a gas at high temperature facilitates the understanding of the quantitative criteria of the forceinteractions of the external flow as well as of the boundary layer. All these problems are of great significance in very highspeed flight during which the heat processes take place fairly intensively. However, the solution of such problems becomes very complicated because they are related to the chemical processes taking place in the gas and also to the chemical interactions between the gaseous medium and the wall material. Keeping in mind the speed range of flight vehicles from low subsonic to very high supersonic flight, the science of the investigation of flow, as indicated in the preface, can be divided into the following fundamental sections: aerodynamics of an incompressible fluid or hydrodynamics (Mach number
6 AERODYNAMICS
M=O); and aerodynamics of high speeds. The latter is further subdivided into the aerodynamics of subsonic (M< l), nearly sonic (transonic M x l ) , supersonic (M > 1) and hypersonic (MB 1) flows. It must be pointed out that in each of these sections the flow interaction processes have some different specific peculiarities. As a result the investigation of each of these flows may be based on different mathematical foundations. Aerodynamic investigations, as already mentioned, are based on the types of flow motions over a bodythe free (external) inviscid flow and the boundary layer flow. The individual sections of aerodynamics deal with each of the types of motions. The aerodynamics of an inviscid (ideal) fluid is attributed to free flows and the aerodynamics of a boundary layer, to boundary layer flows. The aerodynamics of an ideal medium investigates the distribution of inviscid parameters of flowinteraction at the outer limit of the boundary layer. Consequently they are the boundary conditions for the solution of the differential equation in the boundary layer. Normal pressure is related to the inviscid parameters. The resultant forces and moments due to this pressure can be derived from the distribution of these parameters. The aerodynamics of an ideal medium is mainly based on Euler's fundamental equations. Boundary layer science is one of the most widely developed sections of the science of fluid and gas motion. The solution of problems of motion in the boundary layer allows us to determine the distribution of tangential stresses and hence the total aerodynamic forces and moments due to friction. It also helps in estimating heat transfer from a heated gas flow around a surface to the wall of the body. Modern boundary layer theory is based on tlie fundamental investigations of L. Navier, D. Stokes, 0.Reynolds, L. Prandtl and T. Karman. Soviet scientists contributed significantly to the development of boundary layer theory. Academician A.A. Dorodnitsyn worked out the basic theory of a boundary layer in a compressible gas. Professor L.G. Loitsyanskii developed the effective method of calculation of the boundary layer on a curvilinear surface. In the aerodynamic investigations of highspeed flight it is necessary to consider heat processes in the boundary layer in spite of their low intensities. Here the aerodynamics of forces is kept to the fore even though the heat transfer and effect of temperature on the boundary layer should be taken into account at high flight speeds. It is natural that much attention is being paid to the solution of such problems, particularly in these days. In the Soviet Union Professors L.Ye. Kalikhman, I.A. Kibel', V.I. Ievlyev and others are working out gas dynamic theory for heat transfer by investigating viscous flow in a boundary layer over different types of bodies at high temperatures. Similar problems are under investigation by many scientists in other countries.
INTRODUCTION 7
Aerodynamic heating is not the only problem of flowinteractions at hypersonic speeds. The fact that at such speeds an ionization of gas takes place due to the high temperatures and the gas becomes a conductor of electricity raises new problems associated with the control of plasma flow by use of magnetic poles. In the formulation of problems of interactions between the moving body and the plasma the corresponding aerodynamic calculations must consider electromagnetic forces along with gas dynamic forces. These problems are studied in magnetoaerodynamics. A special section of aerodynamics, namely the aerodynamics of a continuous medium, deals with the investigation of fluid and gas motion on the basis of the hypothesis of continuity stated earlier. However, it must be noted that this hypothesis is valid only for the conditions of lowaltitude flight, i.e. in sufficiently dense layers of atmosphere where the length of the mean free paths of molecules of air is small. In the condition of highly rarefied atmosphere at high altitudes this length of the free path of molecules becomes quite appreciable and air cannot be treated as a continuous medium. Hence the conclusions of the aerodynamics of a continuous medium will not remain valid under such conditions. The interaction between the rarefied medium and the body moving through it is studied in a special section of aerodynamics known as aerodynamics of a rarefied medium. The rapid development of this science during the last few years was called for by space research with the help of artificial earth satellites and space rocket vehicles and rocket systems of different types (intercontinental ballistic missiles, global rockets and others) which circle the earth at very high altitudes. The conditions of flowinteraction and consequently the aerodynamic characteristics of an aircraft will vary according to the nature of the variation in the flow parameters at given points in space. The major class of problems of flowinteractions in practice can be solved within the framework of steady aerodynamics, assuming that the flow parameters at given points are independent of time. However, it becomes necessary to account for the unsteady nature of flowinteractions when investigating the stability of flow due to nonuniform flight velocity, oscillations or rotation of flight vehicles, etc. This is due to the fact that under such conditions the surrounding flow will be characterized by the local variation of its parameters with time. Study of this type of flowinteractionsis related to unsteady aerodynamics. We have considered the classification of modern aerodynamics on the basis of types of gas flows. It is obvious that in each of the sections of aerodynamics the investigation of flowinteraction leads in general to different shapes of aircraft or components. Hence it is interesting to consider the classification of modern aerodynamics according to the form of the aircraft or of its individual structural elements. The modern airplane in its aerodynamic scheme basically represents a
AERODYNAMICS
combination of the main body (fuselage), wing and stabilizer and control surfaces. The effects of aerodynamic interferencethe aerodynamic interaction between given elements of the flight vehiclemust be taken into account in the aerodynamic calculation of such bodies. Accordingly the resultant aerodynamic characteristics of lift, drag and moment may be calculated as a summation of similar characteristics of isolated fuselage, wings, stabilizers and controls with corrections introduced as a result of given interactions. Thus a given scheme of aerodynamic calculations requires knowledge of the aerodynamic characteristics of the individual components of the flight vehicle. The aerodynamic calculation of the lifting surface of wings occupies a special section in aerodynamic science known as the aerodynamics of wings. The great Russian engineerscientists N.E. Zhukovskii and S.A. Chaplygin are considered to be the founders of the aerodynamic theory of the wing. The beginning of the 20th century saw N.E. Zhukovskii's discovery of the nature of the lift force on a wing. He invented the formula for calculating it which is commonly known as the Zhukovskii formula. The work of N.E. Zhukovskii on bound vortices representing a hydrodynamic model of the wing is considered to be a very valuable contribution. The series of wing profiles he worked out (Zhukovskii profiles) is widely used in the design of airplanes. Academician S.A. Chaplygin is the author of several pioneer works on the aerodynamics of wings. In the work On Pressure Distribution over Bodies in Rectilinear Flow of 1910 S.A. Chaplygin laid the foundation of the theory of the infinite wing. In 1922 he published a scientific work, Theory of tlte Monoplane in which he gave the theory of a series of profiles (Chaplygin's profiles) and also worked out the theory of the stability of a monoplane wing. S.A. Chaplygin is the originator of the theory of a wing of finite span. The Soviet scientists, Academician A.I. Nekrasov (1 8831954) and Corresponding Member of the Academy of Sciences of the USSR V.V. Golubev (18841954), contributed greatly to the aerodynamics of wings. A.I. Nekrasov worked out basic wing theory in an unsteady flow. V.V. Golubev invented various kinds of high lift devices for a wing, methods of controlling the boundary layer and flowinteraction of smallaspect ratio wings. In his work Interaction of Bodies and Gas at High Subsonic Speeds, Academician S.A. Khristianovich developed an original and very effective method of accounting for compressibility effectsin the interaction of profiles of arbitrary shapes. Abroad, Professor L. Prandtl (Germany) and G. Glauert (England), who also contributed to the approximate theory of thin airfoils in a subsonic flow at small angles of attack, worked on the problems of compressibility effects on the flowinteraction of wings. The results they obtained can be treated as particular cases of the general theory of flowinteraction worked out by S.A. Khristianovich.
INTRODUCTION
The theory of thin wings of various plans in a supersonic flow was developed in the works of Professors Ye.A. Krasil'shchikova and S.V. Fal'kovich. The results of aerodynamic research on wings are applied to the calculation of the aerodynamic characteristics of the stabilizer and also of some control surfaces having plans similar to those of wings. The specific peculiarities of flowinteractions of various types of aerodynamic controls and the presence of other types of components for controls led to the growth of a special section in modern aerodynamics known as the aerodynamics of components of control surfaces. Modern flight vehicles of the rocket type in many cases have the form of rotating bodies or something like them. The combined rocket system of fuselagewingstabilizer has a fuselage (rotating body) as the basic component of the aerodynamic system. As a result the aerodynamics of conical shapes (rotating bodies) in recent years underwent very rapid development as one of the important parts of modern aerodynamic science. Soviet scientists Professors F.I. Franklyand E.I. Karpovich, authors of an interesting scientific work, Gas Dynamics o Thin Bodies, contributed greatly f to the development of the aerodynamics of rotating bodies. A group of scientific workers (K.I. Babenko, G.P. Voskresenskii and others) of the Mathematical Institute of the Academy of Sciences, USSR, worked out the method of threedimensional supersonic flow past sharp bodies, particularly when chemical reactions are taken into account in the flow field surrounding the body. Aerodynamic specialists abroadD. Taylor (England) and 2. Kopal (USA)found the solution of a very important problem in a supersonic flow past a sharp cone. The rapid development of modern mathematics and number techniques in the direction of perfecting the methods of aerodynamic investigations facilitates the solution of many difficult problems of aerodynamics including the problems of aerodynamic interference and corresponding corrections to the aerodynamic characteristics of flight vehicles. The special section of aerodynamic science called the aerodynamics of interference has developed rapidly in recent years. At low supersonic flight speeds aerodynamic heating is comparatively low and cannot cause damage to flight vehicle structures. The basic problem here relates to the selection of the technique of cooling which maintains the required temperature of the wall surface. More complicated problems arise at very high flight velocities of bodies having high kinetic energy. For example, if the flight vehicle has cosmic speed the conversion of even 25 to 30% of this energy into heat is enough to evaporate the vehicle completely. The main problem that arises in the safe launching of space vehicles in dense layers of the atmosphere is the dissipation of this energy so that the minimum amount is absorbed by the body in the form of heat. It was found that bodies with a
blunt nose have such properties. This brought about the development of aerodynamic research on blunt bodies. The Soviet scientists, Academician A.A. Dorodnitsyn, Corresponding Member of the Academy of Sciences, USSR, G.G. Chernii, Professor O.M. Belotserkovskii and others greatly contributed to the study of the problems of the aerodynamics of blunt bodies. Similar investigations were carried out by M. Lighthill (England), P. Garabedyan (USA) and other scientists abroad. The provision of a blunt nose to the body surface can always be treated as an aid in the heat protection of a flight vehicle. In this process the blunt nose itself undergoes the most intensive heat interaction. Therefore this part needs more heat protection than the peripheral part of the apparatus. The most effective heat protection may be obtained by the use of various sheaths of a material that gradually erodes at such temperatures and is carried away bit by bit. In this process a considerable part of the energy imparted to the flight vehicle by the heated air is absorbed. The analysis of theoretical and practical methods for calculating the separated mass (ablation) is related to the modern section of aerodynamic science called aerodynamics of the ablation of surfaces. The wide range of problems of aerodynamics deals in general with the determination of the flowinteraction of the medium with flight vehicles of arbitrary shapes. The forms of flight vehicle surfaces can be selected with the particular aim of fulfilling this or that aerodynamic effect. The form of blunt bodies minimizes the heat transfer to the rest of the body. Consequently the blunt surface is taken to be the optimum from the heat exchange point of view. In designing flight vehicles the problem arises of the selection of shapes producing the least interaction of forces. One such problem is specifically related to the determination of the form of the nose of the flight vehicle that offers least drag at given flight speeds. Similar types of problems are studied in the section of aerodynamics known as aerodynamics of optimum shapes. Such are the contents, basic problems and corresponding classification of modern aerodynamics.
1. ForceInteraction of the Medium on a Moving Body 1.1 Surface force Let us examine the forceinteraction of the medium on a body moving in it, assuming that the medium is a viscous gas (real fluid) and fairly dense. The gas is then taken as a continuous medium, i.e. there is no empty space between gas molecules. The forceinteraction of such a fluid leads to a continuous distribution of forces Fn, due to normal stress, and also to the as distribution of forces FT a result of tangential stresses (Fig. 1.1.1). The resultant force, known as the surface force, acts on the elemental surface dS that we are considering. The vector of this force is determined by the rule of addition of two vectors: F=?in+FT. Here the force E, also includes the additional component derived from frictional force along with the force due to normal pressure. *,.
.
In the case of an ideal fluid it is assumed that viscous forces are absent and that the forceinteraction on the surface of a body produces only the forces that arise from normal pressure. This is quite clear, since the projec11
12
AERODYNAMICS
tion of a force on thissurface exists if the force vector is inclined to the normal at a point on the elemental surface, i.e. a shear stress exists which is however absent in an ideal fluid. The forceinteraction will be the same if a uniform fluid flow over a steady body with the velocity at an infinite distance from the body is equal to the velocity of the body in the above case. The velocity is moreover called the velocity at infinity or velocity o oncoming (undisturbed) flow and is represented f in place of V (velocity vector of the flight). It is obvious by the vector that V=  V . , The oncoming flow is characterized by the undisturbed parameterspressure p,, density p,, temperature T, in place of the corresponding parameters p, p, T of the disturbed flow that exists during the interaction of the body (Fig. 1.1.2). The physical properties of a gas (air) are also characterized by the kinetic parameters: coefficient of viscosity p and coefficient of thermal ,, conductivity A (the corresponding undisturbed parameters will be p, and A) and also by the thermodynamic parameters: specific heats at constant pressure c, (c,,) and at constant volume cv (c,,) and the ratio of specific heats (index of isentropic flow) k = c/, (k, = c,,/c,,). ,c
r,
1.2 Properties of pressure in an ideal fluid Let us examine the properties of pressure in an ideal fluid. Let us first write down the equations of motion of the elementary tetroid MlMoM2M3 of a fluid of sides Ax, Ay, Az along the x, y, z axes respectively (Fig. 1.1.3) by equating the product of the elemental mass and its acceleration to the summation of forces acting on it. These equations can be written down in terms of the components of forces along the axes of coordinates. Let .us restrict ourselves to the formulation of these equations of motion of a tetroid along the x axis, assuming that the other two components will have similar forms. The product of the elemental mass and its acceleration along the x axis is equal to pavA W (dVx/dt),
13
where pa, is the average density of fluid in the elementary volume A W; dVx/dt is the component of acceleration of motion of this mass along the x axis. The forces acting on the mass of M 2 fluid are determined in the following y way: As already mentioned, these forces consist of socalled surface forces. In a given case they are determined by the action of normal pressure on the sides of an elementary fluid mass with projection on the x axis being equal to 0 X X
z pxASx pnASn cos (nx). force, acting On the Fig. 1.1.3. Normal stresses acting on The separated elementary volume of fluid, sides of elemental fluid particle in form of tetroid. is volume (or mass) force which is proportional to the mass of fluid particles in this volume. The universal gravitational forces and, in particular, centrifugal forces are examples of volume forces. The mass force acting on an electrified (ionized) gas in an electromagnetic field is also an example of volume forces due to electromagnetic action. Here the motion of gas under such forces will not be studied (see special course in magnetohydrodynamics). For the case in question the component of mass force along the x axis can be expressed by XpaVA with X as the component of volume force per unit W mass. Considering these values of components of surface and volume forces, the equation of motion will be
A
~ a vW A
dVx
where X is the xcomponent of volume forces per unit mass; ASx and AS,, are the corresponding values of areas MoM2M3 and M I M ~ Mcos (nx) is the ~; cosine of the angle between the normal n to surface M I M z M ~ the x axis; and px, p, are normal pressures acting on the surfaces MoM2M3 and M ~ M z M ~ respectively. Divide the equation so obtained by ASx and, keeping in mind that ASx=
AS, cos (nx), take the limits of Ax, Ay, Az tending to zero. Then terms containing AW/ASx will also tend to zero as A W will have a small value of the third order and ASx is a small quantity of the second order in compariand consequently son of the sides of the elements. As a result (p,p,)=O px =pn. Study of the equations of motion along they and z axes gives,p,=p, and p z =PI,.
14
AERODYNAMICS
As elementary volume is arbitrarily oriented the following conclusion can be drawn from the results so obtained: the pressure at any point in the ideal fluid flow is the same on all the surfaces passing through that point. Consequently, it can be considered as a scalar quantity which depends only on the coordinates of the point and time.
1.3 Effect of viscosity on fluid motion Laminar and turbulent motion: Observations show that a viscous fluid is characterized by two types of motion. The first, laminar or layerwise motion, is characterized by orderly positioned streams that do not intermix. In a laminar flow the transfer of momentum, heat and fluid particles takes place by processes of friction, heat conduction and diffusion of molecules. Such motion is found and is usually steady at small velocities of fluid flow. If the magnitude of the velocities of fluid flow exceeds its critical value at given conditions of interaction of the surface, the laminar motion does not remain steady and becomes a new type of motion characterized by movement of fluid particles from one layer to another. As a result the orderly layerwise flow vanishes. This type of flow is called a turbulent flow. In a turbulent flow the mixing of microscopic particles is superposed on the molecular orderly motion which is typical of a laminar flow. This flow has an unsteady character because the velocity and other parameters at a given point depend on time. In analyzing turbulent flow it is convenient not to deal with instantaneous velocity but with its average value for some interval of time (tz). For example,
the component of velocity along the x axis will be K=(l/tz) J V' dt, where
t13t,
Vxis a component of instantaneous velocity at a given point as a function of and I along axes y and z can , time t. In the same way, the components Fy be written down. Taking the usual meaning of average velocity, the instantaneous velocity can be expressed in the form V, =V,3 Vi, where V; is an additional variable component of velocity known as fluctuating velocity. The perturbation components of velocity along the y and z axes are denoted by V; and V; respectively. The fluctuating component of velocity can be measured by positioning a measuring instrument with small inertia (for example, a thermoanemometer has such properties) at a required point in the flow. The instrument indicates deviation of velocity from the average velocity (i.e. it indicates perturbation velocity) in a turbulent flow. The kinetic energy of the turbulent flow can be determined by summation of kinetic energies, calculated for average and fluctuating velocities. Kinetic energy of the pulsating flow at a point being considered can be obtained as the value proportional to the mean square ofthe perturbation velocity. If the pulsating flow is resolved into components along the axes of co,ordinates the kinetic energies of each of the components of such a flow will be proportional
t1
15

to the respective mean square of the fluctuating components designated by AVi2, AV;, AVi2, which may be determined by
If the equality
is satisfied at a point being considered the turbulence is called isotropic at that point. If this condition is fulfilled at all points it is called a homogenous isotropic turbulent flow. The meaning of average and perturbation values can be extended to pressure and other physical variables. The existence of the perturbing velocities leads to additional normal and shear stresses and to more intensive heat and mass transfer. Frictional stress: Let us examine the formula for frictional force in a laminar flow. In such case the friction appears as a result of diffusion of molecules associated with the transfer of momentum from one layer to another that results in variation of flow velocity, i.e. it leads to the appearance of relative motion of particles of gas in layers. The frictional force, according to the hypothesis first put forward by Newton, is proportional to the value of the variation of the velocity between the mixing particles situated in layers which are unit distance apart from each other. If the distance between the layers is A and the relative velocity of particles Av, the ratio AvlAn in the n limit An+O, for adjacent layers, is equal to the derivatjve avian, known as normal velocity gradient. On the basis of the above hypothesis Newton's formula for frictional stress can be written as
7 =p
(a vlan),
(1.1.1)
where p is the coefficient of proportionality depending on the properties of the fluid, its temperature and pressure. It is also known as the coefficient of dynamic viscosity. The value of this coefficient for gas, by the kinetic theory formula, at a given density p is
and depends on the kinetic properties of gas, like mean free path I and average velocity of motion of molecule This value of p is also called the coefficient of kinetic viscosity. Consider friction in a turbulent flow. Take the simplified model for the presence of additional frictional force in a turbulent regime suggested by
c.
16
AERODYNAMICS
Prandtl for an incompressible medium and the semiempirical relations introduced for it. Now take two layers in a uniform flow characterized by variation of mean velocity only in one direction. Further assume that in one of the layers the velocity V,#O and F,= 0 . Then for an adjacent layer situated at a distance Ay = 1' the mean velocity is equal to V x + ( d E / d y )1'. The particle moving from the first layer to the second, by Prandtl's hypothesis, maintains its velocity V,. Consequently at the time of appearance of this particle in the second layer the pulsating velocity component will be v:= (dvx/dy)1'. It has been found experimentally that away from a solid surface a fluid has properties of isotropic turbulence. So by Prandtl's hypothesis it can be assumed that the absolute value of component V; is proportional to the absolute value of v::
G=
V ; =aV;= al'(dTi,/dy), where a is some constant of proportionality. The absolute value of momentum due to motion of fluid mass p ~ ; d S through elementary area dS will be equal to p ~ (E+ v:) dS. This momen; tum determines the additional force due to stress created by the perturbation components of velocities. From this the absolute value of shear stress is
I ZT I=pv;(Vx+ v;).
Taking the mean value of this expression we get
where V: V i is the mean value of the product of perturbation velocities and 7; is the average value of perturbation velocity. It can be shown that the last one is equal to zero. Integrating the derived quantity Vy= + V ; with respect to t in the interval from t~to ( t l + tz) and then dividing it by 52, we get
tl+
But FY=(l/t2)
,
tl + t z
t2
I1
vy=(I /t2)
tl
17
where coefficient of proportionality a is absorbed in the average value of I' and denoted by I. The magnitude of I is called mixing path length and is, in its own way, an analog of the length of the mean free path of molecules in the kinetic theory of gases. The sign of frictional stress is determined by the sign of velocity gradient; consequently
ZT =p12
I dpx/dy I (dFx/dy).
(1.1.3')
The complete value of frictional stress can be obtained by direct addition of the frictional stresses ZT due to viscosity and due to interchange of molecu= les, i.e. where the value ~ b p (dK/dy) is added to the value of ZT which is obtained from the energy loss of particles in their collision and irregular mixing. Thus z=tb+~~=p (d~x/dy)+p12dvx/dy I (dpx/dy). 1 (1.1.4)
I'
The investigations conducted by Prandtl showed that the path length of intermixing I=rcy, where rc is some constant value. Consequently near the wall (1 1.5) Z ~ = O = Z ~ = p (dvx ldy) I Y=O . I From experimental data it follows that turbulent flow cannot occur in the direct vicinity of a wall, where a very small rate of mixing occurs. Here the flow remains laminar and for that relation (1.1.5) is valid. Under these conditions the stress ~b is very small and hence it can be assumed that the frictional stress is determined by (1.1.3'). Concept of boundary layer: From equations (1.1. I), (1.1.3) it follows that the frictional stress in the different parts of a fluid is not the same for one and the same mediuminteracting body and it is determined by the value of the local velocity gradient. Investigations have proved that the velocity gradient has its highest value near a wall because the viscous medium undergoes stagnation due to adhesion to the surface of the body under interaction. The flow velocity changes from zero at the wall (Fig. 1.1.4) and gradually increases with the increase in the distance from the surface. The frictional stress changes accordingly, i.e. near the wall it is significantly higher than away from the surface. The thin layer of fluid over a surface, characterized by high velocity gradients along normals to the surface and consequently by significantly high frictional stresses, is called the boundary layer. The physical representation of the boundary layer can be understood by covering the surface of a body with colored material soluble in fluid. It is obvious that the paint will diffuse in a layer of fluid and simultaneously will be carried away downstream along with the flow. This colored
zone will represent in itself a layer with gradually increasing thickness along the flow direction. This colored zone of fluid approximately coincides with the boundary layer. It breaks away from the surface in the form of a colored eddy stream (or aerodynamic wake, see Fig. 1.1.4). It is known that with an increase in flow velocity the thickness of the layer diminishes and the wake becomes narrower. The nature of velocity distribution across the section of the boundary layer depends on whether it is laminar or turbulent. The abovementioned distribution of velocity or its mean values with respect to time seems more uniform for a turbulent flow than for a laminar flow (see Fig. 1.1.4) as a result of the cross flow of particles and their collisions. From the distribution of velocities in a flow field near the surface of the body it is found that higher frictional stress exists in a turbulent boundary layer which is usually governed by the increased value of velocity gradients.
Laminar layer
Turbulent layer
Fig. 1.1.4. Scheme of interaction of body and viscous fluid flow: ageneral flow pattern: 1turbulent core; 2laminar sublayer; 3boundary layer; 4surface in flow; 5slip stream (wake); 6free flow; bvelocity profile in boundary layer.
There exists a region of flow within the boundary layer where velocity gradients and hence frictional forces are small. This part of the flow is called the external free flow. This type of flow is also called inviscid. The velocity across the thickness of the boundary layer varies from zero at the wall to a value equal to that in the external free flow. Fluctuating flows, as mentioned earlier, are prevented near the wall and this helps to reduce the fluctuating velocity components. Therefore part of the boundary layer is always laminar. This thin section of boundary layer is called the laminar sublayer. The basic part of the boundary layer outside the laminar sublayer can be either laminar or turbulent depending on the intensity
19
of cross flows in the boundary layer. Thus study of motion in the region of a turbulent boundary layer, called the turbulent core (see Fig. 1.1.4), is at the same time connected with the investigation of fluid flow in the laminar sublayer. The variation in speed across the section of the boundary layer is characterized by the fact that after gradually increasing with the distance from the wall it asymptotically approaches the value of the velocity of the external flow. V , However, for all practical purposes it is convenient to separate the part of the boundary layer where changes in velocity take place quite fast from the upper boundary x where the velocity differs negligibly from its value in the outer free flow. The distance Fig. 1.1.5. Boundary layer scheme: from the wall to this boundary limit is termed Iwall; aouter limit of layer. the boundary layer thickness 6 (Fig. 1.1.5). Usually this thickness determines the point in the boundary layer section where the velocity differs from its value in the outer flow by about one per cent. The introduction of the concept of the boundary layer made it easier to investigate friction and heat transfer because its smaller thickness compared to the size of the body allows us to simplify the differential equations of the fluid flow and hence makes their integration easier.
2. Resultant ForceInteraction
Components of aerodynamic forces and moments: The forces due to normal and shear stresses are continuously distributed over the surface of a body in a fluid flow. They can be expressed by one main vector F o f aerodynamic forces and vector 3 of moment due to these forces about some reference point (Fig. 1.2.1). This point may be an arbitrary point on the body, for example its center of gravity, the nose of a rocket cone, a point on the leading edge of a wing, etc. In practice we do not deal with vectors F a n d @but with their projections on the axes of some system of coordinates. Let us consider two popular systems of coordinates in aerodynamics, nameIy the velocity and the fixed systems of coordinates (see Fig. 1.2.1). In the velocity system of coordinates the longitudinal axis Ox is always directed along the velocity vector of motion of the center of gravity of the apparatus; the vertical axis Oy lies in the plane of symmetry and directed upward (positive direction) and the axis Oz along the span of the right wing (right system of coordinates). For reversed motion the
20
AERODYNAMICS
longitudinal axis coincides with the flow direction and the axis Oz points along the span of the left wing so that the right system of coordinates is maintained. In this case the velocity system of coordinates is also called the flow system of coordinates. In the fixed system of coordinates, rigidly connected with the flight vehicle,, the axis 0x1 is directed along the principal axis of inertia, the vertical axis Oyl lies in the plane of symmetry and the horizontal axis Ozl is directed along the span of the right wing and forms the right system of coordinates. The positive direction of axis Oxl from tail to nose corresponds to the case of the positive direction of motion of the vehicle (see Fig. 1.2.1). In both systems of coordinatesthe velocity (flow) and the fixedthe origin of a system coincides with the center of gravity of the flight vehicle.
Fig. 1.2.1. System of aerodynamic forces and moments acting on aircraft in velocity system (x, y, z) and fixed system (xl,y1, zl) of coordinates.
The projections of vector on the axes x, y, z in the velocity system of coordinates are known as the drag force X, side force Z and lift Y. The corresponding projections of the same vector on axes XI,yl, zl of the fixed system of coordinates are called the axial force (XI or R), the side force Z1 and the normal force (YI or N). The projections of vector M i n both systems of coordinates carry the same names. The component of 2 about the longitudinal axis is called rolling moment (the corresponding symbols in the velocity and fixed systems of coordinates are Mx and Mx, respectively), the component about the vertical axis is called yawing moment (My, My,) and the component about the lateral axis is called pitching moment (MZ, Mz,).
21
The vectors of aerodynamic forces and moments in the velocity and fixed systems of coordinates are then:
respectively, where i, j, k and il, jl, kt are unit vectors along axes corresponding to the velocity and 5xed systems of coordinates. The positive moment about an axis is considered to be the moment which tends to rotate the flight vehicle in a clockwise direction (if the observations are made from the origin of a moment vector). With the accepted systems of coordinates in Fig. 1.2.1 positive pitching moment, for example, increases the angle of attack and negative moment reduces it. The magnitude and direction of action of the forces and moments depend on the orientation of the body with respect to the velocity vector F ( o r if the reverse motion is considered it is relative to the direction of oncoming flow %) at given flight speed and altitude. This orientation of the body is determined by an angle of attack aan angle between the axis 0x1 and the projection of vector 7in plane xlOyland by an angle of side slip pan angle between vector F a n d the plane xlOyl. Transformation of the aerodynamic forces and moments from one system of coordinates to the other: Knowing the angles a and p it is possible to transform, with the help of concepts of analytical geometry, the components of force and moment in one system of coordinates to the corresponding components in the other system of coordinates. The components of aerodynamic forces and moments in the fixed system of coordinates, for example, can be changed to the drag and rolling moment in the velocity system of coordinates by the formulas: X= X1 cos (xlx) + Yl cos Qlx) +Z1 cos (ax);
A A
(1.2.3) (1.2.3')
where cos (xlx), cos (ylx), cos (zlx) are cosines of angles between the axis Ox and the axes 0x1, Oyl, Ozl, respectively.
Fixed system
  
                         Velocity system
OY
TABLE 1.2.1
_______________ox
02
0x1
OYI OZI
sin a cos a 0
22
AERODYNAMICS
Similar expressions can be written for the other components of the force and moment vectors. The values of sines and cosines used for the transformation of forces and moments from one system of coordinates to the other are shown in Table 1.2.1. Using the necessary quantities from Table 1.2.1 in formulas (1.2.3) and (1.2.3') we get X=X1 cosacos/3Ylsinacosp+Zl sinp; (1.2.4) Mx= Mxl cos a cos p My, sin a cos P +Mzl sin P.
(1 2.4')
For example, in the case of motion of the flight vehicle shown in Fig. 1.2.1, equation (1.2.4) gives
X= XI cosacosp YIs i n a c o s p + Z ~ sinp.
Similarly the forces and moments can be transformed from the velocity system to the fixed system of coordinates. For example, the data from Table 1.2.1 lead to the following formulas for longitudinal force and rolling moment: Xl=Xcosacos p+ YsinaZcosasinp; (1.2.5) Mxl=Mx cosacosp+M, sinaMzcosa sinp.
3. Determination of Aerodynamic Forces and Moments from a Given Distribution of Normal Pressure and Shear Stress. Concept of Aerodynamic Coefficients
(12.5')
Aerodynamic forces, moments and their coefficients: Let us examine the following problems: Assume that the distribution of the normal pressurep and the tangential stress z are known on the surface of a body in a flow at a given angle of attack and sideslip under given conditions of undisturbed flow parameters (velocity V static pressurep,, density p, and temperature T,). It is , required to find the resultant values of the aerodynamic forces and moments. The normal force ( p p,) dS and the tangential force zdS due to differential pressure and tangential stresses respectively act on an isolated elementary area dS of the body surface. The summation of the projections of these forces on the x axis in the velocity system of coordinates is given by
A
cos ( E x ) z cos dd~. The other two projections of these forces on y and z axes can be obtained similarly by appropriate changes in cosines. To get the resultant of the forces it is necessary to integrate (1.3.1) over the complete surface S. Introducing terminology for the pressure coefficient p = ( p p,)/q, and the local friction coefficient cfx=z/qwin the above relations, where q, = p,V2,/2 is the velocity head, we get the following expressions for drag, lift and side forces:
[(P P,)
(EX)]
x=:q,s0
(S)
A A dS. [Fcos(F~x)+c~cos(Ex)I
So'
dS
(1.3.4)
The characteristic area Soin the above formulas may be replaced by any given surface like the wing plan area, the maximum cross sectional area of the body, etc. The integrals in (1.3.2) to (1.3.4) are nondimensional quantities comprising the effects of types of interaction of a body of a given geometrical form and the distribution of nondimensional pressure coefficients and friction coefficients of aerodynamic forces. In formula (1.3.2) for the drag force X the nondimensional quantity is usually denoted by cx. It is called an aerodynamic coefficient of drag. In the remaining two formulas the corresponding notations for nondimensional quantities c, and c, are introduced: the first is the coefficient of lift and the second the coefficient of side force. Thus X=cxq,S0, Y=cyq,S0, Z=czqooS0. (1.3.5) The general relations for moments can be obtained in a similar way. For example, consider relations for rolling moment Mx. Obviously the elementary moment dMx is determined by summing all the moments about the x axis arising out of the forces acting on the elementary surface area dS in a direction perpendicular to the x axis. If the coordinates of an elementary surface area dS are y and z from the x axis then the elementary moment will be
Integrating this expression over surface S and introducing the nondimensional parameter .
where L is some characteristic length, we get the formula for total rolling moment, Mn=:inxq Sd. m The parameter mx is called the aerodynamic coefficient moment. (1.3.7)
24
AERODYNAMICS
Similarly the formulas for the other components of moment: My=my 9, SOL, Mz=mz 9, SOL.
(1.3.8)
The nondimensional parameters my and m, are called the coefficients of pitching and yawing moments respectively. The corresponding aerodynamic coefficients of forces and moments can also be introduced in the fixed system of coordinates. With the help of these coefficientsthe forces and moments may be expressed in the following way: XI =R =ex, q So=CR q, So,MXI mxl q, SOL; 7 = I YI =N = cylq, So CN q, So,Myl =my, q, SOL; = (1.3.9) 21 czl gooSo,Mzl =mzl gooSL. = The quantities cx1 (cR), cyl (cN), c,, are called the corresponding coefficients of axial (longitudinal), normal and cross forces and the parameters mx,, my,, m,,, are the coefficients of rolling, pitching and yawing moments. In some cases the above coefficients of moments in the fixed system of coordinates may be represented by
Typical geometrical dimensions: The absolute values of the aerodynamic coefficients, though they happen to be arbitrary to some extent, depend on the choice of the characteristic geometrical dimensions Soand L. However, for the sake of convenience in practical calculations the characteristic geometrical sizes are fixed on in advance in some form or other. Usually in rocket technology the area of maximum cross section of cone S,=S,,, and the length of rocket L are taken as the characteristic area and characteristic length respectively. For the aerodynamic calculations of an airplane system it is customary to use So=Sw, wing plan area, and the wing span I (or the chord b) as the the characteristic area and characteristic length respectively. The line joining the two farthest points of the wing profile is called the chord of the wing. For rectangular wing plan forms the chord is equal to the width of the wing. Usually in practice we find wings with variable chord distribution along the span. For such wings usually the mean geometric chord b =bavor the mean aerody, namic chord b =b are chosen as the characteristic length, where bav=S,"/1 and b,, is the chord of the equivalent rectangular wing with the same area and aerodynamic characteristics. In calculating the forces and moments from the known aerodynamic coefficients one must confine oneself to the geometrical lengths from which these coefficients were derived. If it is desired to carry out calculations on the basis of other geometrical parameters one has to recalculate in advance all the aerodynamic coefficients on the basis of the respective,new geometrical parameters.
The condition of constant values of forces and moments acting on one and the same flight vehicle, while using different geometrical lengths, leads to the relations clSr =czSz (for force coefficients) and mlS1L1=m&Lz (for moment coefficients). Here cl, cz and ml, mz are the force and moment coefficients respectively for two different geometrical parameters SI, and SZ,Lz. LI The respective coefficients c2 and m2 recalculated for the new characteristic parameters S and LZcan therefore be found, since 2
where the earlier dimensions S , L1 and the aerodynamic coefficients cl, rnl 1 and also the new dimensions S2, L2 are supposed to be known quantities. Drag polar of an aircraft: Let us examine the socalled drag polar of a flight vehicle as a practical example of use of the aerodynamic coefficients. It relates lift force to drag or the lift coefficient to the drag coefficients in the velocity system of coordinates. This curve, known as the drag polar of the first type (Fig. 1.3.1, c), represents the geometrical position of the vectors of total aerodynamic force acting on a flight vehicle at various angles of attack (or the vector coefficient CF for this force obtained from the relation CF = F/Soqm).
Fig. 1.3.1. Construction of drag polar of first type for flight apparatus: agraph of function c,=c,(a); bgraph of function c,=c,(a); cdrag polar of first type.
The drag polar of the first type can be constructed from the graphs cx= c,(a) and c, = c,(a) in such a way that the values of the drag and lift coefficients c, and c, are plotted along abscissa and ordinate respectively. Thus each point of this curve corresponds to a particular value of the angle of attack as a parameter of the drag polar. Therefore the indication of angles of attack is shown on the curve c,= f(c,). The drag polar of the first type is convenient in practice as it easily determines, for any angle of attack, an important characteristic of the aerodynamic eficiencv factor.
If the scales for c,(Y) and c,(X) are the same K is equal to the tangent of the angle of inclination between the abscissa and the vector drawn from the origin of the coordinates to the point on the drag polar diagram corresponding to the given angle of attack. From the drag polar it is possible to determine the optimum angle of attack aoptfor the maximum value of the Y / X ratio
if the tangent is drawn from the origin to the drag polar curve. The point c, , on a drag polar curve indicates the maximum lift that , can be achieved at critical angle of attack acr. It is possible to find the point on the drag polar curve showing minimum drag coefficient c, mi, and the respective values of angle of attack and lift coefficient. The drag polar is symmetrical about abscissa if the flight vehicle is symmetrical about the horizontal axis. For such a flight vehicle the value of c, d, occurs at zero lift c, =0 . The drag polar of the second type is also used in place of the above. This drag polar is connected with the fixed system of coordinates. The coefficients of axial force c~ and of normal force CAT are plotted along the abscissa and ordinates of this drag polar respectively (Fig. 1.3.2). This curve is used, in particular, in strength calculations for flight vehicles. Theoretical and experimental investigations show that in most cases the aerodynamic coefficients for the given body shape and angle of attack depend on such nondimensionalparameters as the Mach number M,= V,/a, & Reynolds number Re,= V,Lp,/p,. Here a, is the speed of sound in an undisturbed flow; p, and p, are the corresponding density and coefficient of dynamic viscosity of gas; L is the length of the body. Therefore many drag polar curves exist for each flight vehicle. For example, it is possible to x,) draw a family of curves for the given Reynolds numFig. Drag polar of ber RL. Each one of these curves corresponds to its second type. own value of velocity Mw. The curves on Figs. 1.3.1 and 1.3.2 correspond to the fixed value of Re, and determine the relation of c, and cx in a case where flight takes place at low speeds (of the order of 100 m/sec) at which the aerodynamic coefficients are independent of M,. Center of pressure and aerodynamic center: The point of action of the resultant of distributed aerodynamic forces is called the center pf pressure of the aircraft. It represents some conditional point because the interaction of
lv3e2.
a fluid medium in reality does not lead to a concentrated force but to forces distributed over the surface of a moving body. Usually it is understood that this conditional point lies on one of the main axes of the body, for example the longitudinal axis of the aircraft passing through its c.g., the axis of symmetry of the rotating body or chord of the wing profile. The position of the center of pressure is usually determined by the coordi,, nate x. measured from the nose of the body in a fluid flow. 1f the moment M, about this point and the normal force N (Fig. 1.3.3, a) are known the coordinate of the center of pressure is given as
If the moment M , tending to decrease the angle of attack is considered negative (see Fig. 1.3.3, a) the coordinate x,., is positive. Keeping in mind that
M, = m, q, Sob and N =
we have
Xcep=
CN
qooSo,
mzb/cN,
and hence xc.p/b = CC., = inZ/cN. The ilondimensional value of x,.,, determined as the ratio of the distance of the center of pressure from the nose to the characteristic length of body (in this case, to the wing chord b), is called the coefficient of the center of pressure. At low angles of attack when the coefficients of lift and normal force are approximately equal (C+CN) we get CC., =  mZ/cy. (1.3.12)
The center of pressure, as seen from Fig. 1.3.4, changes its position with the change in angle of attack. This effect is significant at low aniles of attack. In such conditions the coefficient c,., of a cambered profile increases considerably with a decrease in a and becomes infinitely high at the angle of attack corresponding to zero lift. This directly follows from the formula (1.3.12) in which at c,+O the coefficient of moment m, can have sDme finite value m, =~ Z O .
Fig. 1.3.4. Variation in coefficient of center of pressure as a function of angle of attack: 1cambered profile; 2symmetrical profile.
The position of the center of pressure of both types of profiles (cambered and symmetrical), as shown in Fig. 1.3.4, is roughly constant in a fairly large range of angles of attack (from 5 to 20"). The values of c, and m, are simultaneously zero at a+O for a symmetrical airfoil. The following expressions cy=(acy/aa) a, mz= (amz/aa) a (1.3.13) are valid at low angles of attack (here the derivatives acy/aa and am,/aa are constant values which may be obtained at about a zero angle of attack awO) and the coefficient c,., will be equal to some constant value
For flight machines not having horizontal symmetry the flight parameters can be easily understood from the position of the aerodynamic center and not from the position of the center of pressure. To understand the meaning of this take an unsymmetrical profile and calculate moment M,, about an arbitrary point F with coordinate x, lying on the chord of this profile. It is directly seen from Fig. 1.3.3, b that
Mm =N (xn xc.p),
or, since
N . x,.,
0,
we get Mzn =Nxn M.P Keeping in mind the concepts of aerodynamic coefficients and assuming small angles of attack for which c~wc,, we get (1.3.15) m =cy (xnlb) mz. ,
The relation m,=f (c,) assumes a linear form at small angles of attack (Fig. 1.3.5), i.e. (1.3.16) mz=mzo (amdac,) c. ,
Then
where mzo is the coefficient of moment about the leading edge point at c, =0 . The second term in (1.3.17) determines CY the variation in that part of the moment that is connected with the variation in lift coefficient. If a point F on the chord is selected so that the coordinate X n = X F is obtained from the condition (see Fig. 1.3.3, b) xn/b =X F I=XF =  amz/acY, (1.3.18) ~ then the coefficient of moment about this point will not depend on cy and at all (small) angles of attack it will have a constant value. This point is called the aerodynamic center or simply the focus of a given body. Obviously the aerodynamic center represents the point of action of all uniformly distributed forces due to the angle of attack. The relation between the center of pressure and the aerodynamic center is given by ccq,=
CY CY
Fig. 1.3.5. Dependence of moment coefficient on coefficient of lift for unsymmetrical configura6 on of flight vehicle.
where
For symmetrical profiles m,o =O and consequently the center of pressure coincides with the aerodynamic center. Static stability: A chance disturbance (initial throb on starting, a gust of wind, a change in engine functioning from the previous condition, etc.) may change the angles of attack and yaw for a body in motion. After the disturbance disappears the moment arising as a result of this leads to further variations in those angles. If during the subsequent motion the changes in the angles a and P tend to the initial values the flight will be statically stable. If the deviation continues to increase the flight will be statically unstable. Static stability of a flight vehicle is subdivided into longitudinal, directional and lateral static stability. In the case of longitudinal static stability the general longitudinal moment will be stabilized, i.e. will tend to bring back the angle of attack to its initial value. In this case the direction of change of moment M , (and the corresponding coefficient m,) is against the direction of change of the angle a. Thus the condition of longitudinal static stability can be expressed by the inequality aM,/aa<O or am,/aa=m;<O. A positive destabilizing moment which, if it exists, tends to increase the angle of attack is responsible for longitudinal static instability. Therefore the condition of longitudinal static instability will be given by the inequality aM,Iaa>O, i.e. am,/aa=m;>O. Static directional stability is characterized by the inequality a My/a p < 0 or amy/ag=mf < 0; static directional instability by the inequality a Mylap > 0 or amy/ap=m$> 0. In the first case the directions of changes of yawing moment and the angle of yaw are different and in the second case they are equal. Consequently the yawing moments will be stabilizing and destabilizing in the first and the second case respectively. Lateral static stability (or static stability in bank) is represented by the ; derivative a MJ a p or am,/ a/3=m;. If the derivative m < 0 the flight apparatus has lateral static stability. If m > 0 it has lateral static instability. ; The flight apparatus is neutral in the sense of longitudinal, directional and lateral static stability if it has the properties of m:=O, m$=O, mc=O respectively. Use of the concepts of the coefficient of center of pressure and the aerodynamic center in evaluation of longitudinal static stability: The distance between the center of pressure x,., and the center of gravity x,.,, i.e. the difference x,., x,.,, or in nondimensional form (xcap/b) (~c.g/b) ~c.*xcg, = can be taken as the criterion for longitudinal static stability of a flight vehicle. If the coefficient of center of pressure c,., is more than the relative value of the coordinate of the center of gravity x,.,, i.e. if the center of pressure lies behind the center of gravity, then the flight will be statically
31
stable. If the center of pressure is in a forward position (the difference cc.,x,., being negative) there exists static longitudinal instability. The criterion CC.,XC.~ is called the static stability margin. It can be positive, negative or equal to zero depending on static longitudinal instability or neutral longitudinal stability respectively. The concept of the aerodynamic center is used for evaluation of the static stability of unsymmetrical machines or symmetrical flight machines with controls deflected. The coordinate of this point is determined by the formula (1.3.18). Then, assuming the value of x, equal to the coordinate of center of gravity x,., in (1.3.17), we get
from which
This shows that longitudinal static stability depends on the position of the aerodynamic center relating to the center of gravity of the flight vehicle. For a rearward position of the aerodynamic center with respect to the center of gravity (the difference %F % being positive) the derivative amz/ , . acY<0 and the flight will be statically stable. On the other hand if the aerodynamic center is ahead of the center of gravity (the difference xpx,., being negative), then amz/acY>0and consequently the flight will be statically unstable.
4
32
AERODYNAMICS
numbers from M,= V,/a,wO to M,=0.3 to 0.45 (a, is the speed of sound in an undisturbed flow). This idealization of the flow processes assumes that in the given range of speeds the Mach number is taken to be zero because the small disturbances (sonic vibrations) in an incompressible medium propagate at an infinitely high speed and hence the ratio of the flight speed to the sound speed tends to zero. Modern flight vehicles fly at high speeds for which the flowinteraction is associated with a significant change in pressure and hence there exist considerable variations in density and temperature. In highspeed flight conditions it is necessary to take into account the compressibility effect on the flowinteraction with the body as it may be quite appreciable. The compressibility effect of a gas on the aerodynamic characteristics of a body at high speed is one of the important features of highspeed aerodynamics.
4.2 Heating of gas
The rise in flight speeds resulted in the need to investigate the special characteristics of gas flows in aerodynamic research. These characteristics are connected with the physicochemical properties of air. Compressib~lityis taken as the most important characteristic of highspeed gas flows in "usual" supersonic aerodynamics; the effect of high temperatures on thermodynamic parameters, the kinetic coefficients of air and the physicochemical processes occurring in it are neglected. On the other hand at very high (hypersonic) speeds the properties of air related to the effects of high temperatures are given first place. The high temperature appears as a result of stagnation of a gas flow during which the systematic undisturbed flow of gas particles changes to a chaotic motion of the molecules. The kinetic energy of a uniform undisturbed flow is transformed into internal energy of the gas. The excitation of the vibrational modes of the internal energy of the molecules of oxygen and nitrogen in air begins to play an important role at a temperature of about 1,500K. The free vibrational modes of oxygen molecules are found to be completely excited at a temperature of about 3,000"K and pressure of 1 atmosphere. A further increase in temperature makes the atoms break their intramolecular bonds as a result of which, for example, a diatomic molecule breaks into two separate bonds. This process is known as dissociation. Along with dissociation there occurs the phenomenon of recombination, i.e. formation of a new molecule during the collision of two atoms (02+20). This reaction is accompanied by the generation of heat which encourages the collision of the two atoms with a third particle, which carries with it a part of the energy generated and thus helps to create a more stable molecule. Besides this, the chemical reactions taking place in air result in the formation of a certain quantity of nitrogen oxide NO. Further heating leads to the formation of atomized nitrogen and 'oxygen and the
33
The molecules of oxygen are almost completely dissociated at temperatures of the order of 5,0006,000K and pressure of 1 atmosphere. In addition the dissociation of a major part of the nitrogen along with the recombination of atoms in molecules takes place at such temperatures. This process continues according to the equation N222N. The degree of dissociation is determined by the intensity of dissociation, which is equal to the ratio of the number of air particles breaking up during dissociation to the total number of molecules. The degree of dissociation depends on the temperature and pressure. The intensity of dissociation increases with a rise in temperature because the speed and energy of the moving molecules increases with the temperature. This enhances the probability of their collision and disintegration. During this process the rate of dissociation increases with any reduction in pressure (density) as a result of the decrease in the probability of triple collisions of particles leading to the formation of molecules out of atoms. For example, oxygen starts dissociating at T=2,000K if the pressure is equal to 0.001 atmosphere but at normal atmospheric pressure the dissociation of 0 2 begins at T=3,000K. The temperature at which the dissociation of nitrogen starts has a tendency to fall from 6,000K at a pressure of 1 atmosphere to 4,000K at a pressure of 0.001 atmosphere. One more process begins at temperatures of 5,0006,000K as a result of the excitation of electrons due to the large energy flow, which induces the separation of electrons from the atoms of nitrogen and oxygen and also from the molecules of nitrogen oxide. This process is called ionization. Primarily it takes place as a result of the collision of air particles during movements in a heated condition. Hence this kind of ionization is called thermoionization. The process of ionization takes place more intensively with a rise in temperature and it is naturally associated with an increase in the concentration of free electrons. The intensiveness of this process is characterized by the degree of ionization, which is equal to the ratio of ionized atoms (molecules) to their total number. Experiments have shown that nitrogen, for example, is fully thermodynamically ionized (degree of ionization is one) at a temperature of 17,000K and pressure of 1 atmosphere. Variation of specific heats: At high enough temperatures the heat added to air is dissipated not only in terms of an increase in the energy of irrotational and rotational flows of molecules but also in some other processes. These processes are the rise in the vibrational energy of atoms and molecules, the work done to overcome the forces of interactions between the atoms during dissociation of molecules and the separation of electrons from atoms
34
AERODYNAMICS
during ionization. As a result the specific heats increase. The change in the specific heats of air is determined by its temperature until dissociation sets in. For an approximate estimate of the effect of temperature on specific heat at constant pressure the formula is used, where q in its turn depends on temperature (Fig. 1.4.1). For
T > 1,000"K this index may be assumed constant and equal to 0.1. At
T,=28gK the specific heat cp,=0.24 k . cal/kgdegree (in the SI system it is 1,000 J/kg degree). Formula (1.4.1) is applicable up to T= 2,000 to 2,500K, during which the intensity of free vibration reaches the state of excitation. In the first phase of dissociation the x , f l r cp specific heat depends not only on temperature but also on pressure. The calculations of specific heats and also of the ratio of specific heats k=c,/c, under conditions of thermodynamic equilibrium at high temperatures were carried out on electronic computers by a group of Soviet scientists under the guidance of the member of the Soviet Academy of Sciences, Prof. A.S. loflo 2000 3000 4000 T,'X Pryedvodityelyev [17], [la]. These calculaFig. 1.4.1. Variation of indices p, n, tions were conducted for the temperature IC in respective formulas for determirange from 1,000 through 6,000K without nation of specific heats, coefficient consideration of ionization since its effect of dynamic viscosity and coefficient in this interval of temperatures is infiniteof heat transfer. simally small. At very high temperatures the effect of simple ionization in equilibrium was considered. This was assumed to be complete at T= 12,000"K and pressure p =0.001 atmosphere. The data of k=cp/c, at temperatures up to 2,000K and pressures of 1 atmosphere or more are governed by temperature and are practically independent of pressure. The curves drawn from the data of works [17] and [I81 characterizing the variation of cp and k at high temperatures are given in Figs. 1.4.2 and 1.4.3 [7]. Kinetic coefficients: The processes of friction and heat transfer occurring in a viscous heatconducting gas depend on the kinetic coefficients of the gas, such as the coefficients of dynamic viscosity p and heat conductivity A. It is found that in the absence of dissociation the coefficient p depends on temperature only and can be determined by the formula where index n depends on temperature (see Fig. 1.4.1). In approximate calculations the average value of n,w0.7 can be used for quite a wide range of
35
Fig. 1.4.3. Variation of ratio of specific heats at constant pressure and volume (k= c,/c,) for air at high temperatures.
temperatures where the initial value of p may be assumed to be u ,, = 1.82 x kg sec/m2(1.79 x N .sec/m2)at T , =288K. Formula (1.4:2) is applicable up to temperatures of the order of 2,000 through 2,500K. With the rise in temperature the error in this formula increases significantly. Experiments show that at high temperatures up to 9,000"K the coefficient of dynamic viscosity of air under a steady state of dissociation can be estimated within an accuracy of 10% by Sutherland's formula
At temperatures less then 1,500K this formula gives somewhat better results than (1.4.2). More accurate calculations proved that the coefficient of dynamic viscosity at high temperatures also depends on pressure. Fig. 1.4.4 shows the graph [7] characterizing the variation in coefficient p at temperatures up to 12,000K in the pressure range from 0.01 through 100 atmospheres.
Heat conductivity, in the same way as viscosity, does not depend on pressure at temperatures up to about 2,000K and can be estimated by the formula n/n, =(TIT&, (1.4.4) where the index IC in its turn depends on temperature, as seen from Fig. 1.4.1. The average values lcav=0.85 and 1 = 5.53 x , k .cal/m. sec .degree (23.2 wattlm. degree) at T , =26 1O K may be assumed for approximate calculations. The relation of the coefficient of heat conductivity to the temperature and pressure is typical for dissociating air. The corresponding diagram is given in Fig. 1.4.5. The existence of viscous forces, like the process of heat conductivity in gas, is connected with the molecular structure of the substance. During their movements the gas molecules transfer energy and momentum from one place to another. The result of the change of momentum is the viscous force and the transfer of energy leads to the process of heat transfer. From this discussion it is clear that the coefficients of heat conductivity and the dynamic viscosity in gas increase with an increase in temperature. At,the inception of dissociation the nature of variation of A and p will be very complicated. Their
37
5
4
2
1
2000
4000
6000
8000
10000 TPK
values diminish at a low degree of dissociation, which is called for by the losses in internal energy during the process of breaking of molecular bonds and the consequent reduction in the temperature of the gas. With a rise in the degree of dissociation the more intensive breaking up of molecules into atoms leads to an increase in the number of particles that take part in transferring energy and momentum, and thereby elevates the kinetic coefficients. 4.3 Condition of air at high temperatures Equation of state: Experiments on the interaction between an air flow and a body showed that the relations of conventional aerodynamics are quite reliable as long as the air remains comparatively "cold." These relations assume constant thermodynamic characteristics and constant physicochemical structures. "Cold" air is confined within the limits of the hypothesis of constant specific heats and the possibility of applying the thermal equation of state for a perfect gas:
where R and RO are absolute and universal gas constants respectively (Ro= 848 kgmlmol. degree); p and T are density and temperature; (pa& is the average molecular weight of air with constant ingredients; ( ~ l , , ) o is constant. Any gas satisfying equation (1.4.5) is called thermodynamically perfect. The calorical equation of state i=pcp/pR which determines the enthalpy of a gas corresponds to equation (1.4.5). If c p / R = k / ( k  1) is assumed this equation is transformed to
38
AERODYNAMICS
The gas is called calorically perfect if its state is defined by equation (1.4.6) following the condition that c, and c, are constant and independent of temperature. It should be borne in mind that the necessity of considering the variation of specific heats with temperature comes into the picture before the equation of state is applied, this being different from that for a perfect gas. For example, the variation of specific heats with temperature, as shown by calculations, begins at the Mach number of undisturbed flow M,=3 through 4
i .lo!m2/sec2
25
20
15
5
Fig. 1.4.6. iS diagram for dissociating air.
39
during motion across a normal shock wave. At M,=6 through 7 the equation of state for a perfect gas and also the equation of sound speed a2=kRT, (1.4.7) are valid because the components of the heated gas behind the shock wave remain unchanged. Thus a gas is calorically not perfect in this case but it will have the properties of a thermally perfect gas. For a gas with varying heat coefficients the calorical equation of state (1.4.6) gives large errors. The actual relation between the enthalpy and temperature is determined by the more complicated function fi (T) in place of (1.4.6). The equation P =( R o / P ~ v ) pT, (1.4.8) where ,u,=f2 (p, T) is valid for dissociating air having specific heats and molecular weight as the functions of its state. The calorical equation of a dissociating gas medium also has a more complicated nature; in general it can be expressed by i=f3 (p, T). Such a medium does not have the properties of a perfect gas for which an enthalpy depends only on temperature. The nature of thermal and calorical equations of state for a dissociating (real) gas, not following equation (1.4.5), is still more complicated. This is because these equations also take into account the forces of interaction between molecules and the volume of the molecules themselves. These equations can be solved by numerical methods with the help of electronic digital computers and are usually expressed in the form of tables and diagrams of state.
The results of the calculations of parameters for the state of air under thermodynamic equilibrium at high temperatures and under pressures of 0.001 through 1,000 atmospheres are given in the works [17], 1181. On the basis of these results the atlas of diagrams of state was composed by a group of scientists [7]. The famous iS diagram (diagram of enthalpyentropy) of dissociating air for heat calculations in aerodynamics is reproduced in Fig. 1.4.6. In this diagram, showing the calorical equation of state in graphic form, the curves of p=const (isobars), T=const (isothermal) and p=const (isochordotted lines) are drawn. In some cases the diagram of calorical state expressed by the relation of ip with the curves of T=const, p=const and S=const may be more convenient for calculations. This diagram, prepared on the basis of data from the iS diagram, is to be found in work [7]. In the same work there are diagrams of different variants graphically representing thermal equations of state. These graphs, helpful in determining an average molecular weight /I,, for dissociating and ionizing air (Fig. 1.4.7) and also the speed of sound (Fig. 1.4.8) as a function ofp and T [I81 or i and S 171, are important for practical calculations. In studying these graphs it may be noted that the speed of sound changes considerably with temperature and to a lesser extent with pressure. This is explained by the insignificant effect of the composition of air on the nature of the propagation of small disturbances. At the same time any change in the constituents of air during its dissociation considerably affects its molecular weight, which has its own expression in the powerful effects of pressure on the value of pa,. The three characteristics of the section on reduction in pa, depending on temperature may be noted from the curves plotted in Fig. 1.4.7. The first of them is decided by dissociation of oxygen, the second by dissociation of nitrogen and the third by ionization of air components.
5. Basic Relations for a Diatomic Dissociating Gas
5.1 Degree of dissociation Almost always the working medium in aerodynamics is a mixture of gases. Such a mixture in itself is a homogeneous system if the values of all parameters at all points of the system are equal under conditions of thermodynamic equilibrium. At the same time this mixture is not a simple system if the mass components of the different gases, in addition to thc usual parameters of state of gas, are taken into account. The mixture of gases can be taken as a simple perfect gas if its components are inert gases and will not react among themselves. Air under normal atmospheric conditions is such a mixture of NZand 0 2 with some admixture of Ar, CO2, etc. At high temperatures air behaves as a reacting mixture because the
diatomic gases become dissociated and atoms so formed take part in recombination. To simplify each project in terms of models of working medium air is assumed to be a mixture of dissociating components N2 and 02, not reacting with one another. Then each component is assumed to be some conditional diatomic gas consisting of symmetrical molecules of the same type which dissociate into two atoms as a result of double collisions. The atoms can also recombine into lnolecules by triple collisions. This system helps in the study of the mechanism of dissociation of pure dissociating diatomic gas ([46], 1957, No.1). The dissociation is assumed to be uniform. This indicates that during chemical reactions determined, for example by the simple equation for the binary process,
I'D
A2 A+A,
r~
(1.5.1)
the rates of reaction from left to right r~ and right to left r~ (corresponding to the rate of dissociation and recombination) are equal. Investigations of all flows of dissociating medium are connected with the definition of degree of uniform dissociation a. In chemical thermodynamics its value for pure dissociating diatomic gas is given by the relation An interpretation of this equation is found in the work ([44], 1957, N0.5). In equation (1.5.2) the degree of dissociation a is found by the expression where n~ is the number of atoms of element A in a given volume; nA2 is the number of molecules of gas A2 in the same volume. In equation (1.5.2) the quantities pd, Td represent characteristic density and temperature of dissociation respectively. The characteristic temperature Td is determined by the ratio D/k, where D is the energy of dissociation of one molecule of A2; k is the gas constant for a single molecule (Boltzmann's constant). Calculations show that the value of Td does not depend on the temperature of the gas. It is equal to 59,000 and 113,000K for oxygen and nitrogen respectively. The characteristic density of dissociation in general depends on the temperature of the gas. However, the variation of pd in the range of temperatures from l,OOOK through 7,000K for oxygen as well as nitrogen is very small compared to the very high changes in the values of eTdlT in the same range. Therefore it is in practice accepted that the characteristic density pd in this interval can be replaced by its average constant value. Calculations show that in approximate estimates the values of pd= 150 g/cm3 for oxygen and pd= 130 g/cm3for nitrogen may be accepted. With the help of the diagram (Fig. 1.5.1) prepared on the basis of equation (1.5.2) the degree of dissociation for any nondimensional values of T/Td and
43
p/pd can be determined. From an analysis of equation (1.5.2) we arrive at the conclusion that dissociation starts at temperatures much lower than the characteristic temperature. If p/pd= 105 is assumed as the most probable value in the vicinity of the ground then from (1.5.2) or from the graph in Fig. 1.5.1 it follows that the values of a=0.05 (5%) at T I T ~ 0 . 0 5 7 the degree and of dissociation a =0.92 (92%) at T/Td =0.105. For the values of density typical of the upper layers of the atmosphere, for example pdlp = lo7, the ratio of T/Td corresponding to the values of a=0.05 and 0.92 comes down to 0.045 and 0.071.
, b
'5
Fig. 1.5.1. Relation of degree of dissociation of diatomic gas with density and temperature.
Let us consider an equation of state for a gas mixture appearing as a result of the dissociation of diatomic molecules. This equation can be obtained by using the relations for determining pressure p and gas constant R of the mixture of gases, and also the partial pressure component pi:
where p, is the density of component, mi is the mass of the atom or molecule, Ri is the gas constant of component; V is the volume of mixture of gases. Equations (1.5.3) are popularly known as Dalton's law. Now let the component of atom (i= 1) be denoted by an index 'A' and the molecular Since concentration of component ci= component (i= 2) by an index 'My. pilp, where p is the density of the mixture, we get
44
AERODYNAMICS
for atomic and molecular components respectively. and Now from (1.5.3), if the condition m ~ = 2 m A the relation between the masses of xi components pi, p and a are fixed by the usual expressions we get equation of state in the following form:
is where kl(2rn~) the gas constant for 1 gram of component of A2 in the mixture. Multiplying by (1 +a) this formula gives the value of gas constant R (absolute gas constant) for 1 gram of mixture of components A and A2, the ratio of their masses being expressed as a/(l a). Let us assume some conditional value which is determined by characteristic values of density and temperature following the expression
pd
This value is known as the characteristic pressure. Numerical values of are shown in Table 1.5.1. With the help of expression (1.5.5) the equation of state (1.5.4) can be written in more convenient nondimensional form:  p=pT(l+u), (1.5.4')
........................................
Element
TABLE 1 5 1 ..
...................................
pa. lo"',
kg/cm%
Pr d
g/cm3
Td.103,
O K
........................................
Oxygen Nitrogen
ud.107, vd.lo3,
mzlsecz mlsec
Air
___
23 . 41 . 37 .
59 113 100
15 . 34 . 30 .
39 . 58 . 5.4
Mean molecular weight: As is known from thermodynamics, the mean molecular weight of the mixture of gases
where Cj is the mass concentration of an arbitrary component; pi is its molecular weight; the summation sign determines the number of niolecules in the mixture. For diatomic dissociating gas
45
Since
we have
5.3 Thermodynamic relations Let us examine the thermodynamic relations used in the investigation of a diatomic dissociating gas flow. One is the equation of internal energy. It can be obtained from the condition that the internal energy is determined as the sum of the average translational energy of atoms and molecules (312) k T n ~ and also from the rotational and vibrational energy 2(3/2)kTn~, , that the molecules possess additionally. Here the energy of electrons can be neglected. Besides this if the chemical energy of dissociation, equal to (112) D ~ A , taken into account we get an equation for internal energy is
The value of internal energy per unit mass of gas is where V =mA ( n +2 n ~ ~ )isp volume occupied by the gas. ~ l the Introducing the characteristic energy of dissociation per unit mass (per 1 g of A21 Ud = D/2m~, (1.5.9) equation (1.5.8) can be written in nondimensional form
where = u/ud ; values of ud are given in Table 1.5.1. In the same table are shown the values of parameter V =.I&, representing characteristic velod city of dissociation. Physically this parameter represents the velocity that provides kinetic energy to the flow which is equal to half of the energy necessary for complete dissociation of the gas. Comparing (1.5.5) and (1.5.9) and bearing in mind that Td= Dlk, we get the formula
from which it follows that pd characterizes pressure developing during stagnation of the flow, with velocity and density represented by V and pd d respectively. Now let us examine an equation of enthalpy. It is well known that i= u+(p/p). (1.5.11) Dividing terms of this equation by ud, we get the relation for enthalpy in nondimensional form T=U+($;), where i=i/u,. Substituting expression (1.5.8') from equation (1.5.4'), we get the enthalpy relation (1.5.12)
i=a+T(4+a),
If the value of
T= TIT,
based on (1.5.13) we get the relation for the degree of dissociation a as a function of enthalpy ?and the ratio of densities p,/p. This relation is graphically shown in Fig. 1.5.2.
Fig. 1.5.2. Relation of degree of dissociation of diatomic gas with density and enthalpy.
Let us find the enthalpy of dissociation, i.e. quantity of heat necessary for dissociation of a unit mass of gas component at constant p and T. To do this we use the second law of thermodynamics, i.e.
d9 =di  (dplp),
47
This quantity also determines the heat of dissociation iR=ailaa. Differentiat= ?;+1. For the real temperatures ing (1.5.13) with respect to a, we get, at which dissociation takes place the relative temperature T< 1. Consequently the enthalpy of dissociation of the gas component will be
$/aa
5.4 Coefficient of dynamic viscosity Experiments show that Sutherland's modified formula can be used for calculation of the coefficient of dynamic viscosity for a dissociating diatomic gas
Here the concentrations of atomic and molecular gas components are ca=a, respectively. In the absence of dissociation ( c =a =0) the rela~ tion (1.5.16) changes to the usual Sutherland formula. At complete dissociation (a= 1) the quantity in the square brackets in (1.5.16) is equal to 1.42, which shows that the coefficient of viscosity increases by 42% in comparison with an undissociated gas.
CM =(1 a)
5.5 Mixture of diatomic gases Degree of balanced dissociation: So far we have been considering pure dissociating diatomic gas. Strictly speaking these relations are not accurate for calculations of the parameters of gas dynamics and thermodynamics for a mixture of diatomic gases and in particular for estimation of the degree of dissociation. However, since the values of the molecular weight of nitrogen and oxygen are known fairly accurately a modified gas model postulating a mixture of diatomic gases can be reasonably correct. In this context the difference in the characteristic temperatures of nitrogen and oxygen has to be taken into account. Oxygen, with the smaller energy of the atomic bond in its molecules, starts dissociating at comparatively lower temperatures than nitrogen and is found to be completely dissociated when nitrogen is only entering the dissociation phase. If the temperature at the beginning of dissociation of N2 is 3,500K the degree of resulting dissociation of a diatomic model of air at this temperature
will be determined by the relative mass of oxygen in the mixture, i.e. by its value of 0.235. At T< 3,500K the degree of dissociation of the air model is
whqe ao2 is the degree of dissociation of pure oxygen. After completion of dissociation of 0 2 (at temperature T= 3,500K) the value of ao,= 1 and degree of dissociation of air at T>3,500"K will be determined by the expression
a =0.235 +0.765 a ~ , ,
(1.5.18)
where aN2 is the degree of dissociation of pure nitrogen and the quantity 0.765 determines the volume of Nz in air. For completely dissociated air ao2=aN2= 1 and thus the degree of dissociation of mixture a = 1. For approximate calculations of the degree of dissociation a simplified mass composition of undissociated air is taken as 76.5% of N2 and 23.5% of 0 2 , where the 1% of inert gases in the atmosphere is included in the nitrogen. Effective characteristic parameters: The effective characteristic parameters for the dissociation of air model can be determined from the abovementioned mass components of undissociated air. For example, effective characteristic pressure for dissociation
where pm2 pdo2 are characteristic pressures for dissociation of nitrogen and oxygen respectively. Similarly the effective values of other characteristic parameters can be calculated. Values ofpd, pd, Td,ud and V for a given air d model calculated in this way are shown in Table 1.5.1. Enthalpy for uniform dissociation: In the case of the air model in the form of a mixture of nitrogen and oxygen the enthalpy for uniform dissociation is
where the enthalpy of dissociation of component for nitrogen iRN2udNZ = = ( D / ~ ~ A )and for oxygen iRo, =udo, =(D/2m*)o2 (see Table 1.5.1); CN and N, co are the atomic concentrations of nitrogen and oxygen in air respectively (values of concentration can be found in the work [l7]). If air is considered as a diatomic model of gases i~=aud, (1.5.21)
where ud=7.1 x lo3 k.cal/kg according to the data of Table 1.5.1. For a more complicated model of air, representing in itself the mixture of diatomic oxygen and nitrogen, the enthalpy of dissociation can be calculated in the following way: We assume that the whole of the oxygen is dissociated by the time the nitrogen begins to dissociate. Consequently atomic concentra
49
tion is decided only by the breakdown of oxygen, the value of ao2= 1 and aN, = 0. According to this If conditions are such that the dissociation of oxygen takes place partially, the degree of dissociation of the mixture is a< 0.235 and
At high temperatures nitrogen, like oxygen, starts dissociating and enthalpy of dissociation of the mixture will be
For an approximate estimation of the values of i~ we can use tables or a diagrams of the state of air at high temperatures. In this case the enthalpy i at temperature T and pressure p can be found with consideration of dissociation and i can be determined as a difference o where i,=o is enthalpy at the same temperature T but without taking into consideration dissociation. Its value can be calculated by the formula (1.5.24) ia=o =cpw(TIT,)? T, in which for T > 1,OOOK, the index (p =0.1, and at T = 273K specific heat , cp,=0.24 k.cal/kg.deg. The determination of the thermodynamic parameters of air as a model of a pure dissociating gas in the form of a mixture of N2 and 0 2 is limited to the condition that the given gas medium seems to be heated in advance in the initial (undissociated) phase. The adiabatic index k = 1.33 corresponds to the condition of excited vibrational mode of energy. Strictly speaking such a medium does not seem to occur under normal flight conditions. Even then the relations shown above have practical importance. The transition through a shock wave appearing ahead of the body represents itself as an unbalanced process. In practice, it is found that even at very high velocities and temperatures the extent of dissociation is zero immediately behind the wave front but vibrational levels of energy are excited. Thus behind the shock wave the properties of a real and a hypothetical gas coincide. Therefore the accuracy of thermodyilamic and gas dynamic calculations improves with increase in flight speed when the gas behind the shock wave is heated to very high temperatures. Average heat coefficient of mixture: The average heat coefficient of a mixture is given by
where the first term on the right side represents the atomized component of average heat coefficient and the second term represents the molecular component of average heat coefficient. Enthalpy of gas: Enthalpy of the mixture of gas is given by
where s and ii are the concentration and enthalpy of the ith component of the mixture respectively. The value of ii is equal to the total energy including
T
heat energy ( j cp d T ) i of the ith component of the gas and the chemical
0
energy
Enthalpy ii is called the total enthalpy of the ith component and enthalpy i in ( 1 . 5 . 2 7 ) is the total enthalpy of the mixture of gases. If the mixture of interacting atommolecules is considered, keeping in mind that (&hem)*= 0 for molecules, we get
During the flow of a mixture of gases the enthalpy of some of the total volume may be changed as a result of heat flow into or out of it due to heat exchange, transfer of energy directly during inflow and outflow of gas and transfer of energy due to diffusion. The heat of formation (iChe,)i for each component has a constant value. This value for atomized components of nitrogen and oxygen, which occurs as a result of dissociation of air, is equal to the corresponding values of enthalpy of dissociation, i.e.
(ichem)~ U =
~ = iRo O
The main problem of aerodynamic research is the determination of the force of interaction and heat exchange between the fluid medium and the body under interaction. Solution of this problem involves study of the motion of the fluid medium adjacent to the body. From this the variables entering into the governing motion, namely velocity, pressure, density, temperature, etc., are determined for every point in the flow field. Under certain assumptions the investigation can be reduced to the determination of a flow field representing a close link with the velocities of the particles of a fluid, i.e. to the solution of the kinematic problem. Then, from a known velocity distribution the remaining parameters as well as the resultant forces, moments and heat flows are obtained. There are two methods of kinematic investigation of a fluid medium. One is called Lagrange's method and the other Euler's method.
1.1 Lagrangian method The Lagrangian method deals with the motion of the individual particles of a fluid and determines the trajectory for each of them, i.e. their coordinates, as a function of time. Since there are an infinite number of particles it is necessary for this purpose to characterize in a given way the particle which has this trajectory. Thus the coordinates of a particle a, b, c at a given moment t= to designate the particle. This indicates that from an infinite number of trajectories the one that passes through the point with coordinates a, b, c belongs to the given particle. Therefore the equation of the trajectory in a functional form will be as follows:
where a, 5, c, t are called Lagrangian parameters. The quantities a, b, c are parameters determining the trajectory. The components of the velocity vector at any point on the trajectory are equal to the partial derivatives Vx=ax/at, Vy= aylat, Vz= azlat, and components of the acceleration vector are equal to the corresponding second partial derivatives Wx= a2x/at2, Wy= a2y/at2, wZ a2z/at2. =
1.2 Eulerian method The Eulerian method is often used in aerodynamic research. Unlike in Lagrange's method, not the particle of the fluid but a point in space with coordinates x, y, z is considered to be fixed and the variations of velocity with time at this point are investigated. The Eulerian method consists of expressions for velocities of particles of fluid as functions of time t and coordinates x, y, z of a point in space, i.e. in a field of velocities, which are determined by the vector V= Vxi+ Vyj+ Vzk, where i, j, k are unit vectors along axes of coordinates and Vx= dxldt, V, = dyldt, V = dzldt are components of velocity vectors described , by the equations
The quantities x, y, z, t are called Euler's variables. Solving the system of differential equations,
the equations of the family of trajectories in a parametric form coinciding with equations (2.1.1) can be obtained. Here a, b, c are constants of integrations. Taking the complete derivative of the velocity vector with respect to time we get the acceleration vector,
In this expression a?/;lat indicates local acceleration and the remaining three terms give convective acceleration. Projecting vector on axes of coordinates we get the components of total acceleration as follows:
53
The relation (2.1.5) for acceleration corresponds to the flow which is characterized by the variation in velocity with time at a given point and hence the inequality aV7/at#o. This kind of fluid flow is called an unsteady flow. A fluid flow in which velocity and other parameters at a given point are independent of time (ai?/at = 0) is called a steady flow.
1.3 Stream line and streak line At the moment t = to a line can be drawn in a fluid flow so that at every point it coincides with the direction of velocity. This line is known as a stream line. To get a stream line we have to proceed as follows: Let us take some point A1 in a flow at a moment t =to (Fig. 2.1.1, a) and express the velocity of a particle at this point by vector 71. Then take another point A2 in the neighborhood of point A1 and lying on vector vl. Let the velocity vector at this point at time t=to be 7 2 . Similarly we examine point Ag on vector &, the velocity at which is determined at the same moment by vector 7 3 , and so on. As a result of these constructions we get a stepped line consisting of segments of velocity vectors. Reducing these segments to zero and simultaneously increasing their number to infinity we get, at the limit, the line that follows the complete set of velocity vectors. This is the stream line.
Fig. 2.1 . l . Construction of stream line (a) and stream tube (b):
To get the equation of the stream line we make use of the property that at every point on the line the direction of velocity vector 7 must coincide with the vector di=dxi+dyj+dzk, where dx, dy, dz, are projeotions of the element of an arc d i o n a stream line. Consequently vector product d i x V=O, i.e.
54
AERODYNAMICS
if j k dx dy dz =i(V,dyVydz)j(V,dxVx
vx v y
dz)+k(Vy dxVxdy)=O.
v z
Since, for example V,dy Vydz=O, etc. we obtain a system of differential equations: dx/ Vx =dy/ V =dz/ V,. y (2.1.6) So the problem of determining the stream line reduces to the integration of a system of equations (2.1.6). Each of the integrals Fr (x, y, z, CI)=0 and F 2 (x, y, z, C2)=0 of this equation represents a family of surfaces which depend on one parameter C1 or C2. An intersection of these surfaces gives a family of stream lines. In contrast to the stream line, whose formation is calculated for a fixed moment of time, the concept of a streak line is connected with some interval of time during which the particle follows a certain path. From this it follows that the stream line and streak line are the particlepaths of one and the same particle and in a steady flow case they coincide. If the fluid motion is unsteady the stream line and streak line will not coincide. In aerodynamics the concepts of a stream tube and jet of fluid or gas are used. If stream lines are drawn through the points on an elementary closed contour (see Fig. 2.1.1, b) they form a surface which is known as the surface of a stream tube. If streak lines are drawn through points on an elementary closed contour the surface so formed by restricting part of the fluid is called the fluid jet. The stream tube and gas jet, constructed through a point on one and the same closed contour in a steady flow, coincide with each other.
2. Analysis of Motion of Fluid Particles
Investigation of a fluid flow is based on the application of the Lagrangian method with the help of which the motion of a given particle of fluid can be studied and the correct analysis of this motion established. The motion of a solid body is determined by its translational motion or shift and its rotation about an instantaneous axis passing through its center of gravity. Unlike a solid body, the motion of a fluid particle is characterized, in addition, by the presence of deforming motion changing the form of this particle. Let us examine a fluid particle in the form of an elemental parallelepiped with its sides dx, dy, dz. Let us analyze the motion of its side ABCD (Fig. 2.2.1). Since the coordinates of corners of this side are different, the velocities determined at a given moment of time t = to will also be different: VXA VX(x, y), = v~ = v, (X/ dx, y); VYA V (x, Y), VYB V (X dx, y); = y = y v ~ ~ = V x ( ~ + d x , y + d y ) , x~=Vx(~,y+dy); V VYC VY(X dx, Y dy), V ~ = Vy (x, J] +dy), = D
(2.2.1)
55
where x, y are coordinates of point A. Let us resolve the expressions for velocities in Taylor's series by retaining values of only the first order, i.e. the terms containing dx, dy, dz of the first = , y order. Denoting that velocity V ~ A Vx and VA = V at a point A, we get
From these expressions it follows, for example that the velocity component along the x axis at point B differs from its value at point A by the amount (aVx/ax) dx. This indicates that the point B, undergoing gradual change in velocity V in the direction of the x axis, has the velocity (aVx/ax) dx ,
with respect to velocity at point A along the x axis. As a result linear deformation of the segment AB takes place. The rate of this deformation is ex= vx/ax. Point B also moves in the direction of the y axis with velocity V together , with point A and simultaneously rotates about it with velocity aV,,/ax. Similarly it can be established for point D that the relative linear velocity of this point along they axis is (aVy/ay) dy and hence the rate of linear deformation of segment AD will be By = aV,/ay. The angular velocity of rotation of this segment about point A is equal to aVx/ay (the minus sign indicates that point D rotates about point A in a direction opposite to the rotation of point B). As a result of rotation of the element AD and AB a change in angle DAB (Fig. 2.2.2) takes place, i.e. angular deformation of the particle occurs. Simultaneously, the bisector A M of angle DAB rotates and hence some angle dp comes into the picture between it and the bisector AN of deformed angle D'AB'. Thus the particle will undergo rotation along with deformation. Now, the angle of rotation of the bisector (see Fig. 2.2.2) is ~/~=FYz,
where'
The angles daz and dal shown in Fig. 2.2.2 are equal to (aVy/ax) dt and ( a V,/ay) d respectively. Hence t
Therefore the angular velocity dp/dt of the rotation of a fluid particle about the z axis can be obtained. Denoting it by w,, we can write d =(21 av, avx). "dt p ax= If the motion of an edge AD relative to segment AB is considered it is clear that the angular velocity of this sector is 2 ~= ( a v x / a ~ ) vYiax). , +(a The quantity
is known as the semirate of deformation of the right angle DAB. Let us extend all these results to a threedimensional flow and examine point C, occupying a particle in the form of an elemental parallelepiped with lengths of sides dx, dy, dz. The velocity at this point at time t= to is a function of coordinates x +dx, y +dy, z +dz. Expressing velocity components in the form of Taylor's series and retaining terms only of the first order we get:
Vzc= Vz+
a vz dx+dy+avZ ax a~
av, dz. i
a2
We introduce notations similar to those in the analysis of the motion of a plane particle. Assume that 8,=aVz/az. This value determines the rate of linear deformation of a threedimensional particle along the z axis. Now we introduce the notations:
The quantities cox and o,, represent the components of angular velocities of
57
this particle about the x and y axes. The components of angular velocity my, w, of a particle are taken as positive if rotation takes place from the x axis to the y axis, from the y axis to the z a& and from the z axis to the x axis respectively. Accordingly, the signs of derivatives 8 V,/ax, aV,/ay, aVx/az will coincide with signs of angular velocities and the signs of derivatives aVx/ay,aVy/az, aVz/ax will be opposite to those of the angular velocities. By analogy to (2.2.4') we write the expressions:
COX,
which are equal to the semirate of deformation of two right angles of a parallelepiped in planes yOz and xOz respectively. With the help of simple manipulations it can be shown that
avz/aJ'=Ex+COx;
~ V ~ / ~ Z = E , 
Considering these expressions, the componentsof velocity at a point C (2.2.5)can be written as follows:
Vxc= Vx+8x d x f ey d ~ + ~ Y z C O ,~  0 2 dy; ~ + d (2.2.8) VYc=Vy+Oyd y + & , dx+ex d z + o , d x  o x dz; V , c = V Z + e Zd z + c x d y + e y d x + o x d y  o , d x . Thus the motion of point C can be taken as the result of summation of three types of motion: translational motion along the trajectory together with point A at a velocity V ( v x ,Vy, V,), rotation about it with an angular velocity
and deforming motion. This concludes the essence of Helmholtz's theorem. Deforming motion, in its way, is built up from linear deformation characterized by coefficients ex,OY, 8, and angular deformation which is represented in terms of E XcY, E,. , The linear deformation of the edges of an element limits the variation of its volume z =dxdydz, which is determined by the difference
where dx,, d y l , dzl are lengths of sides of the element at time t + d t . Substituting the values of side lengths, we get
dz = (dx +8 , dx dt) (dy + 8, dy dt) (dz + 8, dz dt)  dxdydz.
Neglecting terms higher than of the fourth order in this expression, we have
From this, the rate of change of relative volume or rate of deformation of specific volume 0 =(117) dtldt can be determined as
which is equal to the sum of rates of linear deformation along any three mutually perpendicular directions at every point in the flow. The quantity 0 is also known as the divergence of the velocity vector at a given point, i.e. div
V=e=ox+ey+eZ.
(2.2.11)
So it is clear that the motion of a fluid particle is of a complex nature and that it can be expressed as the summation of three types of motion: translational, rotational and deforming. A flow in which the fluid undergoes rotation is called a rotational flow and the components of the angular velocity of rotations OX,coy, coZare called the vortex components. To identify rotational flow the concept of rotation of velocity rot v , represented by rot v=2& used. The rotation of velocity is represents by itself a vector rot F=(rot and the components (rot F)x= 2wX,(rot v), =2wy and (rot
(2.2.12)
V),= 2wz
are equal to twice the corresponding values of the vortex components. 3. Irrotational Motion of a Fluid In investigating fluid motioil its rotation is often not taken into account due to the infinitesimally small angular velocities of the particles. This kind of motion is called irrotational motion. For irrotational flow ;=O (or curl F=o) and hence the vortex components wx, wy, wt are zero. Accordingly, formulas (2.2.3), (2.2.6) give (2.3.1) vx/ay = v,lax, v,/a~ av,/az, avx/az= vzlax. = These equalities are the necessary and sufficient conditions for the sum of the three terms (V,dx+ Vydy+ Vzdz) to be a total derivative of some function which characterizes fluid flow in the same way as the velocities Vx, Vy, Vr. Denoting this function as p (x, y, z, t) at any time t, it can be written as dp= V dx+ Vy dy+ V dz. x z
On the other hand the same differential will also be dp =< a ~ / a x ) + ( a ~ ~ ady)+( a ~ i a zd.. dx y ) Equating the last two expressions we get
v, =apiax, V, = a p ~ a y vz= a ~ ~ a z . ,
(2.3.2)
59
This function p is called the velocity potential or potential function of an irrotational flow. The existence of this function characterizes the flow as a potential flow. It follows from the relation (2.3.2) that the partial derivative of function p with respect to any coordinate is equal to the projection of velocity on the respective axis of coordinates. This property of potential function is maintained even for any arbitrary direction. In particular, the tangential component of velocity at any point on an arbitrary curve s will be equal to the partial derivative V,=ap/as and the normal component will be V, = aplan, where n is normal to the arc s at the point being considered. The projections of velocity vector T i n polar coordinates r and 6' at some point in the directions of the polar radius vector and the perpendicular to this radius vector will be the partial derivatives
V=grad p,
(2.3.3)
V= grad p = (grad
(2.3.4)
where the coefficients in brackets on the right side indicate the projections of the vector of the velocity gradient on the coordinate axes: (grad p)x = Vx = aplax, (grad ply= VY a p / a ~ , = (2.3.5) (grad p), = Vz aplaz. The use of a potential function considerably simplifies the study of fluid motion because in place of investigating three unknowns such as the components of velocity Vn, Vu, VZ, it is enough to find one unknown function p. Then, with the help of this function, the velocity field can be fully determined. 4. Equation of Continuity
4.1 General form of equation of continuity The equation of continuity of motion in mathematical forin represents by itself the law of conservation of massone of the fundamental laws of
physics. This equation is related to a number of basic equations in aerodynamics which are used to find the parameters governing the motion of a gas medium. Let us take some moving volume of fluid and obtain the equation of continuity. This volume, consisting of the same particles, varies with time. According to the law of conservation of mass it can be written that p,,z = const for this volume, where pa, is the average density of the fluid contained in the volume z. Hence the derivative d (pavr)/dt O or, seeing that the density and = volume are both variable quantities,
This equation deals with any arbitrary finite volume. To get the relation characterizing fluid motion at a point we take the limit of r+O in the above equation. This indicates contraction of the volume z to some inner point. If it is assumed that the moving fluid completely fills the volume we are considering, i.e. if there is no vacuum or discontinuity, then density p at a given point will be completely determined and can be expressed by the equation
indicated in (2.2.10)and (2.2.1 1). The equation (2.4.1) represents the equation of continuity in general form.
4.2 Cartesian system of coordinates Let us study the equation of continuity in the Cartesian system of coordinates. This could be done by finding the derivative dp (x, y, z, t)/dt and replacing div in (2.4.1) by (2.2.1 1). The continuity equation will then be obtained in the following form:
pv
 a = V) div ( p V ) ( p z + a ( p v y ) + a ( p v Z ) ax
a~
a.
we get
(2.4.3) ( a p ~ a t ) div ( P V )=0 + in place of equation (2.4.2). The equation of continuity was obtained by assuming that the moving volume is filled with the same fluid particles. In a given case we studied fluid motion using Lagrange's method. The same equation could be obtained from
61
the analysis of the flow of different fluid particles through some fixed volume by the application of Euler's method. The difference between the fluid flow rate from the control volume and the momentum of the fluid entering this volume during a given interval of time, on the assumption of continuity of medium, is equal to the change in momentum of the fluid in the given volume in the same interval of time. Let us put this condition in mathematical form. Take a control volume in the form of an elementary parallelepiped having sidesdx, dy, dz. Let the fluid mass fx= pV,dydzdt flow from the left side of the volume in time dt. The mass efflux during this time through the right side is equal to fA+(afx/ax) dx. Consequently the difference between the mass efflux and the influx along the x axis will be
Similarly the expressions for the difference between the mass outflow and inflow along they and z axes could be written as
The change in mass of the fluid in the control volume occurs due to the change in density of the fluid. At time t the mass of fluid in control volume is pz and at time t +dt is equal to pz+(ap/at) zdt. That is, change in mass is equal to (ap/at)zdt. By the condition of continuity this quantity is equal to the sum (af,/ax) dx+(af,/ay) dy+(af,/az) dz, which leads to equation (2.4.2) obtained above. The continuity equation (2.4.2) describes an unsteady flow. For a steady flow we have ap/at = 0 and hence
or
v)
(2.4.4')
or (2.4.7) div V= 0. The continuity equation for potential flow, taking into account (2.3.2), can be put in the following form:
62
AERODYNAMICS
The equation so obtained is known as Laplace's equation. It is known that any harmonic function could be the solution of this equation. Therefore the velocity potential q in an inY compressible flow represents such a harmonic function.
Formulas of transformation: dynamic problems can be conveniently solved by x using a system of orthogonal Fig. 2.4.1. Elementary particle of fluid in curvilinear coordinates. The cylindrical system of coordinates. cylindrical and the spherical systems of coordinates, in particular, are such curvilinear systems of coordinates. In the cylindrical system of coordinates the position of any point P in space (Fig. 2.4.1) is determined by angle y between the reference plane and the plane passing through point P and coordinate axis Ox, and the rectangular coordinates x and r in this plane. The formulas of transformation from the Cartesian coordinates to the cylindrical ones take the following form: x=x; y = r cos y; z = r sin y. (2.4.9) In a spherical system of coordinates the position of point P (Fig. 2.4.2) is determined by the angular coordinates 0 (polar distance), yl (azimuthal angle) and the polar radius r. The relation between rectangular and spherical coordinates is given by the following relations: X = ~ C O Sy=rsin0coscy, ~, z=r sin B sin cy. (2.4.10) The corresponding transformations of the above aerodynamic equations from Cartesian coordinates to one or other system of orthogonal curvilinear coordinates Fig. 2.4.2. Elementary particle of fluid in may be achieved in two ways: spherical system ofcoordinates.
either by direct substitution of (2.4.9), (2.4.10) in these equations or by application of a more general method based on the concepts of generalized coordinate curves [8]. Let us examine the latter method. The elementary arclengths of coordinatecurves about a point P can be written in the following general form: ds1= hl dqi , d ~ =h2 dq2, d ~ =h3 dq3, 2 3 where q, (n = 1, 2, 3) are curvilinear coordinates; hi coefficients are known as Lame's paraaeters. Then we have for cylindrical coordinates (2.4.1 I)
From Fig. 2.4.1 it directly follows that for cylindrical coordinates dsl =dx, dsz = dr, ds3 = rdy. (2.4.14)
From Fig. 2.4.2 the arclengths of corresponding coordinate lines in a spherical system of coordinates will be
Lame's parameters in cylindrical coordinates will then be: and for spherical coordinates they will be:
Let us examine some expressions for vector and scalar quantities in curvilinear coordinates which will be necessary for transformation of the continuity equation in these systems of coordinates. The gradient of some scalar function @ is grad @=asl
a@il+iz+aa,
as2
aa, i3
ass
where il, i2, i3 are unit vectors along the corresponding coordinate lines. From formulas (2.4.1 I) it follows that grad @=.it+. h, aql hz aq2
a@
a@,+.1
For further transformations the material of a course in mathematics on vector analysis may be used. The divergence of the velocity vector can be obtained by expressing it as a sum of components along coordinate lines
2
n=l
(2.4.20)
The following wellknown relation of vector analysis is used in determination of div in, div in= div (i, x ij) = ij rot im imrot ij, (2.4.21)
where n takes the successive values of n = 1, 2, 3 corresponding to the values of m=2, 3, 1 and j= 3, 1, 2. In this relation there appear the vectors rot im and rot i. for which the general representation of rot in can be introduced. Keeping (2.4.18) in mind and using general methods of transformation of vector quantities, we find
Now we substitute (2.4.22) in (2.4.21). Naturally, for this substitution it is necessary to change index n in rot in to indices m and j and also to effect corresponding changes of indices on the right side of (2.4.22). Then we arrive at
Now it is possible to substitute these equations in (2.4.20). Considering in grad V,,= 1 .aVn , hn aqn from (2.4.18) and using it in the above expressions, we get the following relation for divergence:
 1 div V= hi h2 h3[
(2.4.24)
aql
aq2
a43
Keeping this expression in mind, the transformation of continuity equation (2.4. l ) to different forms of orthogonal curvilinear coordinates can be discussed. Equation of continuity in cylindrical coordinates: In cylindrical coordinates Lame's parameters are given by the relations (2.4.16) and the values of qn (n = 1, 2, 3) are given by relations (2.4.12). The velocity components along the axes of cylindrical coordinates are:
Vl = Vx=dxldt,
V2= V =drldt, r
(2.4.25)
Taking this into consideration, the divergence of the velocity vector in (2.4.24) will be
wheref (x, y, z, t) is a function in Cartesian coordinates and may be used for transformation of derivative dp/dt in equation of continuity (2.4.1) to the cylindrical coordinates. Then
Introducing values of div V and dp/dt in equation (2.4. l), one gets
For potential flow, the following relations could be substituted in the equation of continuity:
vx=aplax,
vr=aqlar, V = (llr) . ( a ~ l a ~ ) . Y
(2.4.251)
If the motion of the gas is axisymmetric, like the flow past a rotating body at zero angle of attack, the flow parameters do not depend on the angular coordinates y. The continuity equation will then have a still simpler form:
Equations (2.4.5) and (2.4.32) can be combined into one equation by writing
where E =0 for twodimensional plane motion and E = 1, y =r for twodirnensional axisymmetric flow.
66
AERODYNAMICS
Spherical coordinates: Using relation (2.4.24) and taking into account formulas (2.4.13) for spherical coordinates, relation (2.4.17) for Lame's parameters and the expressions for velocity components
dr 1 de V1=Vr=& V2=Ve=.7 V3== =r sin 0 Vr r dt dt
9 3
(2.4.33)
The total derivative of density dp/dt, taking into account equation (2.4.27), could be represented in the following form:
ap v V+ ap aP+V,+.+. e a p at ar r a6 r sin O ay' Using values of div P from (2.4.34) and the derivative dpldt from (2.4.35) into (2.4.1) and grouping the terms properly, we will have
ae
(2.4.37)
sine
ay
In the case of an incompressible fluid (p=const) the above expression reduces to 1 v 2)~ . a ( ar , . r a n 8 . a ( v 6 s i n e )'' 10 av@o. sin a y r 88 (2.4.38)
The following changes are used for transforming the continuity equation for the case of potential motion:
vr=a~ %,
1 Vo=. ar r 80'
, , =1.
aP r sin 8 8y'
4.4 Equation of continuity for gas flow over curvilinear surface Let us examine the particular form of continuity equation in orthogonal curvilinear coordinates that is applied in the study of a flow over a curvilinear wall. The x axis, in this system, coincides with the wall contour and they axis is directed along the normal to this wall at a selected point. The coordinates of point P in a plane (Fig. 2.4.3) are equal to length x measured along
the wall and distance y along the normal to the wall respectively. Assume that the wall is a surface of rotation interacted by an axisymmetric flow of gas. The curvilinear coordinates of point P will be The elemental arclengths of coordinate lines are
d s = [I +(y/R)] dx, dsz=dy, ~ ds3 =rdy. (2.4.41)
where r. is the radial coordinate of point P measured along the normal to the axis of the rotating surface; R is the radius of curvature of the surface at a given cross section. Then Lame's parameters will be
h ~ =+(ylR), 1 h3= r. h2= 1, (2.4.42)
Fig. 2.4.3. Equation of continuity in curvilinear coordinates.
Now we use formula (2.4.24) for div where the velocity components are
I div V= 
a cvxr) , ax
(2.4.44)
+ ap+ ,
1 +x
Y
vy
ap.
a~
(2.4.45)
Incorporating expressions (2.4.44) and (2.4.45) in (2.4.1), the equation of continuity, after some modifications, will be obtained in the following form:
In'investigating a gas flow near a wall with curvature in a thin layer on the surface (for example, in the boundary layer), the coordinate y g R. Consequently we have
The equation so obtained is similar to that for a surface with no curvature. In the case of a steady flow we get
For a flow near a plane curvilinear surface the continuity equation will have the following form:
Thus for a flow near a wall the continuity equation in curvilinear coordinates will be the same as in the case of Cartesian coordinates. For a steady flow ap/at=O, so the equation of continuity will be
where E = 0 for twodimensional plane motion and E = 1 for twodimensional axisymmetric flow.
4.5 Equation of mass flow Consider the particular form of continuity equation for a steady fluid flow in the form of a jet. The mass of fluid in some fked volume enclosed by the jet surface and the plane cross sections does not change with time because at each point in the flow the condition of ap/at=O is satisfied. Then the quantity of fluid entering the volume across a plane cross section of area SI per unit time will be equal to ~ I V I S IThis is the same as the mass of fluid p2V2S2 . flowing out through the opposite section with area S (pl, p2 are densities and 2 VI, V are velocities at the first and second cross sections of the jet respective2 ly). Thus PI VIS1=p2VzSz6 As this equality can be applied, to any section it
69
can be written in general form as p VS= const. This equation is known as an equation of mass flow.
5. Stream Function
The study of irrotational gas flow can be simplified if it is connected with the solution of only one unknown potential function which fully determines this flow (see Section 3). It will be shown that for certain types of rotational flow there exists a function that also determines its kinetic characteristics. Let us take a twodimensional (plane or axisymmetric) steady rotational flow of fluid. From continuity equation (2.4.32') it can be shown that some function I,U of coordinates x, y exists that can be determined by the relations:
~ I , U I =Y VX, a y ~ a = pyCv,. ~ PY' x (2.5.1) Then from (2.4.32') it follows that a2t+v/ayax=a2cy/axay,i.e. this identity is obtained. Substituting (2.5.1) in the equation of stream line V,/dy = Vx/dx, which can be expressed in the form
we get
From this it follows that equation (2.5.2) represents the differential of function I,U and hence
~ I , U 0. =
(2.5.3)
(x, y) =const.
(2.5.4)
The function cy, known as a stream function, completely determines the velocity of rotational flow by the relations
It should be noted that it is necessary to take E = 0 for a plane flow and E = 1, y = r for a threedimensional axisymmetric flow. The family of stream lines for a potential flow can also be characterized by a function ly=const, which is related to velocity potential by relations
Assuming p=const, in (2.5.5) and (2.5.6), we get the corresponding expressions for an incompressible flow:
Assuming E = 0 in these last equations we get equations for an incompressible plane flow: Knowing the velocity potential, the stream function can be determined from these equations with an accuracy to the arbitrary constant and vice versa. In a potential flow the family of equipotential lines (in a plane flow) or equipotential surfaces (in an axisymmetric flow) can be drawn besides the stream lines. This family of lines or surfaces can be determined by the equation rp = const. Let us examine the vectors
arp gradrp=if ax
a~ j
a~
a ~ and grad^=i+J, ax
a~
a~
whose directions coincide with the directions of normals to the curves y = const and ly= const respectively. 'The scalar product of these vectors is
Taking into account formulas (2.3.2) and (2.5.5), it can be shown that this scalar product is equal to zero. It follows that stream lines will be orthogonal to equipotential lines (in a plane flow) or to equipotential surfaces (in an axisymmetric flow).
6. Vortex Lines
A curve S drawn in a flow field at a given moment of time such that at each point the angular velocity vector 6coincides with the tangent to this curve at that point is called a vortex line. A vortex line is developed in a similar way to a stream line (see Fig. 2.1.1). The difference between them is that in place of uniform velocity vectors the  angular velocity vectors (WI,co2, etc.) for rotation of particles are considered for vortex lines. So now the vector product x d$=0, i.e.
71
Keeping in mind that, for example, wy dzo,dy=O, etc., we get the equation of the vortex line from the above condition
A vortex tube can be obtained by drawing vortexlines through the points on an elementary contour with cross sectional area o. The product of wo is called the intensity or strength of the vortex tube or of the vortex. We will now prove that the intensity of the vortex tube is constant at all its sections. For this purpose we will use the analogy of an incompressible fluid flow. For the case of incompressible fluid motion it was shown that div V=O.The mass flow rate of jet VISI= VzS2= ...=VS=constant is an example of this equation. Let us study the vortex motion and the expression for divergence of the angular velocity vector,
Substituting here the values of components of from (2.2.3), (2.2.6) we get the expression div &=o under the condition that the second derivatives of of Vx, Vy, Vr should be continuous. Using the analogy of the equation of mass flow rate VS=const, we get the equation of flow for vector; along a vortex tube in the form This equation represents Helmholtz's theorem that the vortex strength remains constant along the vortex tube. From this theorem there follows the property of a vortex tube that it cannot suddenly terminate or end in a point. The latter is based on the fact that as the area of the cross section of a vortex tube 00 the angular velocity w tends to infinity by Helmholtz's theorem, which is physically impossible. 7. Circulation of Velocity
7.1 Concept of the circulation of velocity The concept of the circulation of velocity is very important in aerodynamics. It is used to investigate the flow interaction of an aircraft and, in particular, to determine the lift force acting on a wing. Let us consider some known closed contour K in a fluid flow. Assume that the velocity 7 at each point on this contour is known. Let us now find the integral along this contour
where. T. 7; is the scalar product of two vectors V and d i The quantity r . calculated in this way is known as circulation of velocity around a closed contour. As V,i+ Vyj+ V,k and di=dxi+dyj+dzk, the above integral may be written as
v=
r = vxdx + vY + V, dz. dy
(K)
 A
(2.7.2)
Now, taking into account that the scalar product Vdi= V cos (Vds) ds= Vsds, where V is the projection of velocity on the tangent to the given con, tour at a given point, then
r= vsd ~ .
(K)
(2.7.3)
If the contour coincides with the stream line of circular form of radius r, at each point of which the velocity Vs is constant in magnitude and direction, then r= 27crVs. (2.7.4) The circulation of velocity in an irrotational flow can be expressed'in terms of velocity potential. Assuming that VxdxiVydy+ Vzdz= dp, we get, for an open contour L,
r
where qAand q, are the values of the potential function at the ends of the contour. For a closed contour pA and p, represent the values of velocity potential at points A and B on the contour which coincide with one another. If the potential function is singlevalued qA= p, and the circulation of velocity around the closed contour is equal to zero; a multivalued velocity potential (pA#p,) gives a value of circulation different from zero. 7.2 Stokes' theorem Let us consider an elementary contour ABCD (see Fig. 2.2.1) and find the circulation around this contour. Assume that the velocities are constant along each side and are given by the following values: dy (CD), Vy (AD). ay Taking anticlockwise circumvention of the contour as positive (from the x axis toward the y axis), we get for circulation
vx
~ J J
+a vydx (BC),
vx
+=
73
d r z=
According to (2.2.3) the value in brackets is equal to double the value of the component of angular velocity about the z axis, i.e. 2wz. Consequently
d r z = 2 o,dxdy.
(2.7.6)
The products of differentials in these expressions represent areas contained in the corresponding elementary contours. Taking into account the results obtained above for an area da arbitrarily oriented in space and restricted within an elementary contour, the circulation may be given by (2.7.7) where onis the component of the angular velocity about the direction of normal n to the elementary area da. From(2.7.7) it can be seen that the circulation of velocity around any closed contour is equal to twice the value of the intensity of vorticity lying inside the contour. The relation (2.7.7) can be extended to the case of contour L of finite size having some surface area S for which the value onis known at each point. Then the circulation of velocity around the contour is
dm =2 &do,
Formula (2.7.8) represents Stokes' theorem: the circulation of velocity around a given closed contour L is equal to twice the value of the integral of intensities of vortices through the surface enclosed in the contour. The quantity on the righthand side is called the vortex strength and is designated by
If the circulation r in (2.7.8) is replaced by the formula (2.7.3) we get a relation representing the integral along the contour K by an integral over the surface S enclosed by the same contour K,
The above relations are given for a singly connected contour (the region being considered is limited by one contour, Fig. 2.7.1, a). However, Stokes' theorem can also be extended to multiconnected contours (the region is surrounded by one external and some internal contours). Here equation (2.7.8")
74
AERODYNAMICS
is applicable under the conditions that the external contour is connected with the internal contours by auxiliary lines (section) so that the singlyconnected region may be obtained. Then the double integral in (2.7.8') can be extended to the shaded region (Fig. 2.7.1, b) and the line integral is taken around the singlyconnected region so obtained, i.e. around the external contour along all sections and also along all internal contours. In accordance with this and by formula (2.7.8") and also from Fig. 2.7.1, b, where a triplyconnected region is shown, we get
from which the circulation around contour K for the given region will be
Fig. 2.7.1. Singly connected and triplyconnected regions in plane: asinglyconnected region (TKcirculation along contour K); btriplyconnected region (TKcirculation along external conalong internal contour K1and K2 tour; r K 1 and TK~circulations respectively).
7.3 Velocities induced by vortices The vortices appearing in a fluid create velocities in the surrounding space. This effect is similar to the electromagnetic effect of a conductor carrying electric current. Velocities created by a vortex are called induced velocities. The electromagnetic analogy above acknowledges the fact that the BiotSavait formula, used to determine velocity induced by a vortex, is similar to that representing the law of electromagnetic induction established by BiotSavart.
75
Let us consider a curvilinear vortex of arbitrary form (Fig. 2.7.2). The velocity vector d i induced at a point A by an element of vortex dZ is determined by the radius vector 7 and coincides with the direction of vector ; product 7 x dE, i.e. the vector d is perpendicular to the plane containing vectors F and dZ. The value of dG is determined by means of the BiotSavart formula which has the following form in vector notations:
where r is circulation of velocity, equal to double the value of the intensity of the vortex. Conclusions of formula (2.7.9) can be found in the book [23]. Since I x dI; = r sin a dL, where a is the angle between the direction of an . element of the vortex and the radius vector & the value of induced velocity at a point A is given by
Let us apply the BiotSavart formula for calculation of the velocity induced by the section of line vortex (Fig. 2.7.3). Since r =hlsin a, dL = rdalsin a = hda/sin2 a, we have sin a da = (COS  cos az). a1 4zh
4zh
For a vortex with both ends extending to infinity (infinite vortex) a1 =0,a2 = n and consequently
w=
r
2nh
(2.7.12)
For a vortex with one end extending to infinity and the other at the origin point A (semiinfinite vortex) a1 = 0 , a2 = n/2. Therefore
If there are two or more vortices in a fluid they interact with each other. As a result the vortex system will be in motion. The velocities of this motion are determined with the help of the BiotSavart formula. Consider, as an
76
AERODYNAMICS
example, two infinite vortices having equal intensity and direction of rotation and (Fig. 2.7.4, a). These vortices impart velocities V2=  (r/27~h) Vi = r/271h to each other, which are equal in magnitude and opposite in direction. As a result of this both vortices will rotate a) about an axis passing through the midpoint between them. If one of the two vortices has an intensity of vortex opposite in sign (Fig. 2.7.4, b) the induced velocities will be in the same direction and consequently the system of b) vortices will be displaced uniformly with velocity V=T/2nh in a direction perpendicular to the line joining these vortices. The motion of an incompressible potential flow can be completely determined if the potential function p or stream function cy is known. The relation between these two functions is given by equations (2.5.9), which are known as the CauchyRiemann equations in the theory of functions of complex variables. These equations represent the necessary and sufficient conditions such that the combination of the two functions p + iy happens to be the analytical function of complex variable z =x Iiy, i.e. the functions can be differentiated at all points of the given domain. Let us introduce the symbol for this function: W(z)=p+iy. (2.8.1) The function W (z), which can be determined if the functions of the two real variables p = p (x, y) and cy = cy (x, y) satisfy the differential equations (2.5.9), is called the complex potential. If the values of the function p(x, v) or ~ ( xy) , are known uniquely in the flow field the complete twodimensional plane flow
Fig. 2.7.4. Interaction between vortices: avortices with same direction of rotation; 6with opposite directions of rotation.
8. Complex Potential
77
can be obtained by the complex potential. Hence the problem of estimating such a flow can be reduced to the determination of function W(z). Let us find the derivative of function W(z) with respect to the complex variable z:
This expression is known as the complex velocity, the modulus of which gives the velocity itself, i.e. Id W/dzJ= 4V + V,2_ V. It is obvious that the actual ; velocity vector V= Vx+iVy happens to be the reflection about the x axis of the vector of complex velocity. Denoting the angle between the vector dW/dz and the x axis by 9, the , velocity components will be Vx= V cos 6 and V = V sin 9. Using Euler's formula cos 9  i sin 9 = ej8, we get dW/dz = Veie. (2.8.4)
9. Typical Fluid Flows
Let us consider simple incompressible fluid flows, their geometrical pattern (aerodynamic specter), expressions for complex potentials and the corresponding potential functions and stream functions. 9.1 Uniform flow Let the fluid motion be given by a complex potential W(z)=V(cos 9isin 6) z, (2.9.1) where V and 6 are some quantities constant under given conditions. In accordance with (2.8.1) q+iiy=V(cos 9isin 9) (x+iy), from which the velocity potential and stream function are obtained: a,=V(xc0~9+ysin9);
(2.9.2)
ty=V(ycos9xsin9). (2.9.3) It follows from the expressions for y, and ly that the given flow is plane and steady because time does not appear here as a parameter. The stream lines and the streak lines coincide in such a flow. The velocity components of flow can be found from (2.9.2): aplax = V, = v cos 9, splay = vy= v sin 9, av/az = V, =O. (2.9.4)
Here Y is the resultant velocity of flow and B is the angle between its direction and the x axis. Equating stream function y/ (2.9.3) with constant, including in it the velocity V, we get the equation
y cos B x sin B = const,
(2.9.5)
from which it can be seen that the stream lines themselves represent the parallel lines inclined to the x axis by an angle B (Fig. 2.9.1). Since the velocity components Vx and V are positive the flow direction will be that shown , in Fig. 2.9.1. This kind of flow is known as plane uniform flow. In a particular case where the flow is parallel to the x axis (8 = 0, V,= V, V, = 0) the complex potential will be W(z)=VZ. (2.9.6)
9.2 Plane source and sink Consider the complex potential W(z) = (q/2n) ln z,
Fig. 2.9.1. Steady uniform flow.
(2.9.7)
where q is a constant quantity. This equation can be rewritten in the form r + iy =(q/2n) In (rei8) (q/2n) (In r + iB), p = where r is the distance to the point with coordinates x, y (polar radius), 0 is the polar angle. From the equation obtained above it follows that
Ir/ = (9/2n)B. From (2.9.8) we find that the radial component of velocity (along radius r) is
(2.9.9)
The flow so obtained consists of stream lines (trajectories) representing a family of straight lines passing through the origin of coordinates (this follows from the equation of stream line y= const). This radial flow starting from
79
the origin of coordinates is called the plane source (Fig. 2.9.2). The fluid flow rate across a contour of radius r will be equal to 2zr Vr= Q. Substituting here the value of T / , from (2.9.10), we get Q=q. Conkequently the constant q is determined by the fluid flow rate of the source. This quantity q is known as the strength or intensity of source. There exists a fluid flow, along with the source, called the plane sink. The complex potential of sink is W (z) =  (9127~) z. In (2.9.1 I ) The minus sign indicates that unlike a source flow the fluid flow for sink will be toward the center. Sink, in the same way as source, is characterized by strength or intensity q (flow rate in unit time).
9.3 Threedimensional source (sink) Threedimensional source (sink) occurs in a similar way to plane source (sink). The flow from such a source (sink) is given by the following conditions:
where R = 2/x2+y2 +z2; q is the intensity of source (plus sign) or sink (minus sign). The intensity of source (sink) is equal to the quantity q determined as the flow rate per second through a spherical surface of radius R. The total velocity is
and it coincides with the direction of radius vector R. Therefore the velocity potential depends only on R and can be written as
one gets
y,
M sin 8 2n r
Assuming cy=const and taking into account that r= .r/x2+ y2 and sin 8= y/r =y/.fx2 +y2, the equation of the family of stream lines for a given flow will be y/(x2+y2)= const. (2.9.1 8)
The family of stream lines represents an infinite series of circles each passing through the origin of coordinates with centers on they axis (Fig. 2.9.3, a).
Fig. 2.9.3. Determination of dipole: astream lines of dipole; bscheme of formation of dipole.
To get the physical picture we assume a flow resulting from the superposition of flows from a source and a sink of equal strength situated on the x axis at a very small distance E from the origin of coordinates (Fig. 2.9.3, b). The velocity potential at a point M (x, y) due to flow from source at a distance rl will be qsourc,=(q/2~) r l . Similarly the velocity potential at point M due ln to sink at a distance rz from this point will be y, sink= ( 9127~) r2. In The principle of superposition of incompressible flows can be used to determine the resultant flow from a source and a sink. The velocity potential @ink. In line with the of the resultant flow by this principle y = y,,ur,,+
81
equation of continuity (Laplace equation) the following relation is obtained from (2.4.8'):
the value of a2q/ax2 a2rp/ay2is also uniformly equal to zero. Consequently + the resultant function a, satisfies the equation of continuity. The resultant potential due to source and sink, then, is
As r ~ = d ( x + e ) ~ + y ~ ,= d ( x   e ) ~ + y 2 , get ~2 we
The value of E can be selected so that the second term in brackets will be small compared to unity. Applying the formula for expansion of a logarithmic series and neglecting terms of higher order, we get
Let the source and sink come closer (e+O) and at the same time their intensities increase in such a way that the product 9.28, in the limit of source and sink approaching each other, tends to some finite quantity M. The combined flow appearing as a result of this is called the dipole. The quantity M characterizing this flow is called the strength of the dipole and the x axis, the axis of the dipole. In equation (2.9.19), for y, using the limit of E+O and 2ge+M, one can get the expression for the dipole,
which coincides with (2.9.16). So the flow under study, characterized by the complex potential (2.9.15), happens to be for dipole. The same thing can be shown if we consider the stream function of a mixed flow that coincides with (2.9.17). It can easily be seen that the formulas (2.9.16) and (2.9.17) can also be expressed in the form:
l,y=

a In
sin 0.
(2.9.17')
Let us take a threedimensional case. As in (2.9.14), the potential functioil for the flow created by the source and sink of equal strength q, which are ' placed on axis O (r = dy2+z2)at points r =  E and r = + E, will be r
If we take the limit of c0 and assume that the product 9.28 tends to a finite quantity M, the velocity potential from the above expression for a flow created by dipole with strength M will be
If the source and sink of equal strength q are located on axis Ox at the points x= E and x = +E, then at E+O the velocity potential of dipole is
9.5 Flow with circulation (vortex) Let us study the flow given by a complex potential
From (2.9.23) it is found that radial component of velocity V,.=ap/ar=O and the normal component V, perpendicular to the radius is equal to the derivative of p with respect to arc length s along the stream line, i.e. The equation of stream line can be obtained from the condition of I,V= const which, according to (2.9.24), gives the equation r=const. c his equation itself represents a family of stream lines in the form of concentric circles. Along these circles the direction of motion is positive if it is anticlockwise (from axis x to y), in which case the coefficient a in (2.9.25) will have the plus sign. This kind of flow with particles moving (circulating) along concentric circles is called a flow with circulation or vortex flow (Fig. 2.9.4). The circulation of velocity in the flow being considered has r = 2 n r (aplas). Since aplas =(l/r)aplae =a/r the circulation r = 2 n a or a = r / 2 n . Thus the physical meaning of the constant a lies in the fact that its value is determined by the circulation of flow which, as established earlier, is equal to the strength of the vortex. The flow created by a vortex lying at the origin of coordinates where V, =a/r+w is called the plane vortex source or simply the vortex. We have examined Fig. 2.9.4. Circulating f o (vortex). lw cases of flow for which the velocity potential and stream function are exactly determined. Combining these flows it is possible, under given conditions, to determine more complicated potential flows equivalent to those arising during a flow past bodies of given shapes.
The equations of motion constituting the theoretical foundations of aerodynamics form part of the group of basic equations of aerodynamics. They connect the variables determining motion such as speed, normal and tangential stresses, etc. The solution of equations of motion helps to determine these unknown quantities. Let us examine the different types of equations of motion useful in the study of gas flows. 1.1 Cartesian coordinates Let us study the motion of fluid particles having the form of elementary parallelepipeds of sides dx, dy, dz about a point A with coordinates x, y, z. The velocity components at this point are given by V V,, V,. The motion of , fluid particles with mass pz (z=dxdydz, elemental volume) takes place under the action of the mass and surface forces. The components of mass force are represented by Xpz, Ypz, Zpz, and the components of surface force by Pxz, P,z, P,z. The values of P,, P,, P, themselves represent the projections of vectors of surface force counted per unit volume of body. Equations of particle motion in projection on the x axis will have the form pz dVx/dt= Xpz +Px 2, from which dVx/dt =X+ (11~) Px, (3.1.1) where dVx/dt is total acceleration along the x axis. Similarly, two more equations for projections on axes y and z can be obtained: dVy/dt= Y+ (11~) Py; (3.1.2) dV,/dt=Z+(l/p) P,. (3.1.3) The surface force can be expressed in terms of stresses acting on the sides
84
85
of an elemental parallelepiped. The peculiarity of the surface forces in comparison with the ideal (inviscid) medium lies in the fact that not only normal but also tangential stresses will be acting on the surface. Every surface force acting on the side will have three components along coordinate axes (Fig. 3.1.1). The projections of surface forces acting on unit area of the lefthand side are p,, z .c,,. The quantityp, represents normal , stress and zx,, Z X ~ tangential stresses. It is obvious that the first index are shows the axis perpendicular to the side being considered and the second index indicates the axis along which the given stress is projected. The components of stress p zzx z,,, act on the given side perpendicular to the , z axis and the components pyx, z,,, zzyact on the lower side perpendicular to they axis. Here the normal stresses are assumed to be positive if they are directed away from the separated element and place the body under stress as shown in Fig. 3.1 .l. The tangential stresses will have the plus sign if these stresses are oriented along directions opposite to the positive directions of the axes of coordinates for three sides intersecting at the starting point A. For the other three sides they are positive if directed toward the positive directions of these axes. Having noted this we will study the components of surface forces along the x axis. The force due to normal stress pxx dydz acts on the left side and the force [px, + (ap,/ax) dx] .dydz due to normal stress acts on the right side. Consequently the resultant of these forces is equal to (ap,/ax)dydzdx. The components of forces as a result of tangential stresses acting on these sides will be equal to zero. It is, however, necessary to consider the tangential stresses z and zyx. , Force zzX dydx acts on the rear face and force [zzr+ (azZx/az)dz] dxdy on the front face. The resultant of these forces is equal to (az,,/az) dxdydz. Similarly the resultant of the forces acting on the lower and upper faces will be equal to (a~yJay)dxdydz.
In this way the projection of surface force on the x axis taken for unit volume is Similarly the projections of surface force on other coordinate axes per unit volume are:
Tangential stresses acting over orthogonal sides are equal to each other, i.e. zzx=T.*z, zzy=zYz and zy, =zxp This can be proved by writing the equations of moments about the axes acting on the particle as a result of forces which are perpendicular to the sides and passing through the center of the parallelepiped. For example, the equation of moments of forces about an axis perpendicular to the front and rear faces will be as following: zyx dxdz (dy12)  [zyx (azYxlay) d~dz dyl (dy/2) zxy dydz (dxI2) [zxy (a~~,/ax) dydz (dx/2) =AM,, dxl where AM, is the moment of inertial forces during relative motion of the element about this axis. In this equation, the moments of gravitational forces and normal components of surface forces do not exist because it is assumed that the resultant of these forces passes through the center of the parallelepiped. i It is worth noting that the inertial force, by nature, is the volume force proportional to the elemental volume z=dxdydz and is consequently in this context a small quantity of the third order. Therefore the inertial moment obtained by multiplication of this force by an infinitesimally small arm will be a small quantity of the fourth order and hence can be neglected. Neglecting other small values of the fourth order in the equation we get an equality zxy=zyx. Similarly the other two equalities of tangential stresses can be proved. In this equality of tangential stresses, lying in two mutually perpendicular elemental areas passing through some point and acting in the plane normal to both the elemental areas, the property of mutual dependence of tangential stresses is included. Thus only three tangential stresses will be unknown out of the six. The law of proportionality of stresses to strains is used to determine the values of tangential stresses. Newton's formula for determination of friction arising during motion of a viscous fluid relative to a solid wall is an illustration of the application of this law. According to this formula zyx= p(a Vy/ax), i.e. the stress proportional to semispeed e, =(112) (avY/ax)of angular deformation in the direction of the z axis, from which we may write * ~ y * = 2 pThis. relation can be extended for the general case of spatial ~~ motion in which the angular deformation in the direction of the z axis is
+ +
87
determined by semispeed of deformation E Z = (112) (aVy/ax Vx/ay). The other two values of frictional stresses can be written in the form ~,,=2,ue,, T ~ ~ = ~ InUthis ~ . , E way:
TYX
+a
(3.1.5)
The abovementioned law of proportionality can be used to establish the relation for normal stresses p,, pyy, pzz. Under the action of stress p, the fluid particle experiences linear deformation along the x axis. 1f 8 repre: sents the value of linear strain, it can be stated that pXx=E B ~ ,where E is the coeRcient of proportionality or modulus of elasticity of the fluid. The particle will also be deformed in the directions of axes y and z under the action of stresses pyy and p,,, which reduces the deformation along x axis. It will be recalled from courses on the strength of materials that the reduction of this strain,for elastic bodies, i.e. the quantity d o x , is proportional to the sum of strains along the y and z axes under the action of given stresses. Accordingly
where m is a constant quantity, known as the coefficient of transverse linear deformation. The complete strain along x axis will therefore be Similarly the strains 8, and & along axes y and z can be calculated. The normal stresses can be found from the expressions so obtained:
P X X = E 8x+(lIm) ( p y Y + ~ z z ) ;
(3.1.7) pYY=E & + ( I 1 4 (pzZ+p*); pzz=E &z+(llm) ( P X ~ + P Y Y ) . The relative linear deformations of a particle along the directions of coordinate axes determine its relative volumetric deformation. Expressing the value of this deformation by 87 we can write B=8,+By+ez. (3.1.8)
Performing summation of (3.1.7) and considering an expression for 8, we get mE Pxx +pyy +pzz= m  2 6 Determining the sum of pyy +pzz from this expression and substituting it in (3.1.6), we find m m E PXX =  8 , +  ~ 0m+l m + l m2
88
AERODYNAMICS
The known relation between the modulus of rigidity G, the modulus of linear elasticity E and the coefficient of transverse deformation can be used for the last derivations. This relation is valid for elastic media, including compressible fluids:
G =mE/2 (m + 1).
(3.1.11)
It has been established that for an inviscid fluid the pressure p at any point in the flow (see Chapter 1, Section 1) is equal for all the elemental areas passing through this point, i.e. in terms of established notations p,=p,,=p,,= p. Consequently, p = ( 113) (pxx+pH +pzz). So in investigating the motion of a viscous fluid the pressure can be determined as the arithmetic mean, with negative sign, of the three normal stresses corresponding to three mutually perpendicular planes. ~ccordingly p. The following relations for shear (tangential) stresses acting on a solid body are established in the theory of elasticity:
a=
where yyx, yYz, yzx are angular strains about axes z, x, y respectively. It can be seen from a comparison of formula (3.1.15) with the corresponding relations for a viscous fluid (3.1.5) that they can be obtained from each other if the modulus of rigidity G is replaced by coefficient of dynamic viscosity p and angular strains y by the corresponding values of angular deformation
rates E . Accordingly, in replacing G by p in formulas (3.1.14), (3.1.14') the linear strainsg;, 8,, 8, must be replaced by the corresponding values of the rates of linear deformation ex= aVx/ax, 0, = aV,/ay, eZ= aVz/az (see Section 2 of Chapter 2) and the rate of volumetric deformation
( 2 %$ div T ) ;
The second terms on the right side of expressions (3.1.16) give additional stresses due to viscosity. Using the relations (3.1.4), (3.1.4'), (3.1.5) and (3.1.16), one can find P,, P, and P,. For example, from (3.1.4) the value of P, will be
where A is the Laplace operator: A = a2/ax2 a2/ay2 a2/az2. In particular, Similarly the relations for P, and P, can be obtained. The relations for the stresses due to surface forces in a fluid were obtained here by applying the generalized rules connecting stresses and deformations in solid bodies to the case of a fluid medium having the properties of elasticity and viscosity. The same relations can be obtained from the series of hypothetical statements about molecular forces acting in the fluid itself (see [212],[81). Writing the components of t ~ t aacceleration by the law of calculating the l derivative of multivariable function f (x, y, z, t ) (x, y, z are also functions of
time t), the equations of motion (3.1.1) to (3.1.3) can be obtained in the following form, taking into account the above expressions for Px, Py, PI:
where v =,u/p, coefficient of kinematic viscosity. The differential equations (3.1.17) form the theoretical basis for the gas dynamics of a viscous compressible fluid medium and are known as the NavierStokes equations. It is assumed in these equations that the coefficient of viscosity ,LL is a function of coordinates x, y, z, i.e. ,u=f (x, y, z). Assuming that p=const, the NavierStokes equations can be written in the following form:
The effect of mass forces need not be taken into account in the investigation of gas flows. Consequently, we take X= Y = Z = 0 In that case .
For twodimensional plane motion with variable values of coefficient of dynamic viscosity (,u#const):
If div V=Oand /'=const these equations lead to the equations of a viscous incompressible fluid. Let us examine the equation of motion of an ideal (inviscid) compressible fluid. Taking the coefficients / L an v as equal to zero in (3.1.17), we get
These equations were first obtained by Leonard Euler and are known as Euler's equations. They form the theoretical foundation of the science of the motion of an ideal gas, the hypothetical properties of which are determined by absence or negligibly small effect of tangential stresses.
1.2 Equation of motion in vector form Multiplying equations (3.1.18) by unit i, j, k, respectively and summing them, we get the equation of motion in vector form: dV  1 (3.1.21) =Ggradp+v A V +L grad div 7 , dt 3 P
where the vectir of mass forces in Cartesian coordinates gradient of pressure grad P = (aplax) i +( 8 ~ 1 8j~ )(aplaz) k; + vectors
A ~ = A V ~ j+AVZ k; i+AVy  a div 7 a div 7 a div 7 j kgrad div V= i+  + ax a2
a~
92
AERODYNAMICS
V=Oand consequently
7.
Taking (3.1.21") and (3.1.22) together, the equation of motion can be written in the form
+grad
at
a7
vZ +rot
1 8=   grad p + v ~ + V :
P
grad div
V. (3.1 22')
1.3 Curvilinear coordinates Let us transform equation (3.1.22') with the help of the concept of generalized curvilinear coordinates qn discussed in Section 4 of Chapter 2. This helps us to carry out changes comparatively easily in the equation of motion pertaining to the particular form of curvilinear orthogonal coordinates in the same way as in the equation of continuity. Let us consider the transformation of individual terms in (3.1.22'). The following expressions, bearing in mind (2.4.18), represent the second term on the left side and also the first and third term on the right side of (3.1.22'):
grad p =
C
3
2
3
in ;
(3.1.24)
grad div
8
(grad div
)., in=
a div V i,.
(3.1.25)
For the transformation of vector product rot Vx Vit is necessary to find the formula for vector rot T i n generalized coordinates. To do this, find the rotation of both parts of equation (2.4.19):
93
Here the components of vector grad V, in the corresponding coordinates are determined from 'expression (2.4.18), where @ is replaced by Vn. Inserting the values of (3.1.27) and the expressions for rot in from (2.4.22) in (3.1.26), the rotation of velocity will be
where the component of this vector on a tangent to the corresponding coordinate line is
Note that the indices m = 2, 3, 1 and j = 3, 1, 2 correspond to the values of n= I, 2, 3.  The left side of equation (3.1.22) represents vector of total acceleration W =dF/dt which can be written as
where W nis a component of the acceleration vector along the direction tangential to the coordinate lines g,. It is obvious that each value of Wncan be taken as the sum of the corresponding components of vectors aF/at and of vectors (3.1.23) and (3.1.29) along the given directions. Accordingly
1
n=l
(3.1.33)
94
AERODYNAMICS
where dVn is the component of velocity vector along coordinate lines q,. The first 'vector on the right side of (3.1.33) has already been determined in (3.1.25) and the second vector can be obtained with the help of equation (3.1.28). Performing rotation of both parts of this equality, we get rot rot 7 = y ( r o t rot
3
Y 1, in
(3.1.34)
where the corresponding components of the vector of rot V are found from (3.1.28). With these data the transformation of the equation can be studied with respect to the particular form of curvilinear orthogonal coordinates. Cylindrical coordinates: Using (2.4.12), (2.4.16) and (2.4.25), we get ql=X, Hence q2=r, q3=7, hl=l, h2=l, v, = v,, V = v,, 2 v3= vy. hs=r,
(grad P)I =aplax. Moreover we find that (grad div 61 div v a x . = Keeping in mind that div v i s determined by the formula (2.4.26), we get
av,
a.
(rot V)2=
:t.:(.r
1
7) avx ~[~(~~g)]{
a7
95
a2
a2 vx
1 a2vx 1 + r ~ay2
+u ap.
'
as;
'
So the equation of motion for the component along the x axis in a cylindrical system of coordinates will be
ax 3 ax In the same way the other two equations for the components along coordinate lines r and y wilil be:
P
ax
ar
a?
I I
i
(3.1.35'j
a div Y
J
I
Divergence of velocity is obtained by formula (2.4.26). For an axisymmetric flow the equations of motion are simplified so that
ax
ar
ax
. adivv. 3
a x It (3.1.36)
!@+v
ar
Avr+. a div  I 3 ar J
v,lat = avrlat = v,lat =o. these equations. Spherical coordinates: Spherical coordinates, Lame's parameters and components of velocity vectors along the directions of coordinate lines are related by the formulas (2.4.13), (2.4.17) and (2.4.33): q ~ = r , 42=e, q3=y, h ~ = l , h2=r, hs=r sin 0, V1=Vr, Vz=Ve, v3=v,. Using these details, equation (3.1.32) gives the component of acceleration in the direction of coordinate line r: at ar r a0 + r sin 0 a y The pressure gradient component will be (grad P)I = aplar. Further, using (2.4.34) for div F w e get the relations (grad div V)I = = 1 'rsin~
v, 
avr v$+vi
r
(3.1.37)
(3.1.38)
ae
1 an 0
.a%]e
ay
This relation is used to determine the component (AV), of vector AT by formula (3.1.33). Before doing this we find the value of (rot rot pj from (3.1.34) which occurs in (3.1.33). From (3.1.28) we have: (rot v)3=;

1 a(Ver) 8Vr

[;i;1 as
a
'
1 [ a Vr (r sin 0 VI) (rot V)2=   rsine ay ar Substitution of these relations in (3.1.34) gives
a (r sin 0 V,)
ar ay
e(+r$!
( v a r +ae~ cos o ~ ) ~ ~
g: Fi )a;
sin e
97
 2
r2 sin 0 a y
av, .
Using relations (3.1.37) to (3.1.39) and (3.1.411),the equation of motion for the component along coordinate line r will be
a+v,.+v r aVr
at
+V
Ve .+, aVr V, aVr vE+vf =.1 ar Y ae r sin e at,v r P 2 ..2 aV, 2Vr 2 V ~ e c o t 0 AVr. aVe r2 as r2 sin e a p r2
ap ar a div V
The other two equations, in the same way, will be: +Vr+ ave at ar
we
vo.+.+v, avo
Y
ae
avo
ay
sin 0
V r V ~  V : c o t e =   1 ap . r rp a0
1
I
= 
g)
?e ( sine(0) a
+r,sin20.~e
(3.1.44)
98
AERODYNAMICS
The equations of motion can be written in still simpler form in the case of twodimensional gas flows with variations in variables (velocity, pressure, density, etc.) along two coordinate lines only:
where
(3.1.47)
Equations of twodimensional motion of gas along curved surface: For the case we are considering the curvilinear coordinates, Lame's parameters and velocity components are determined by equations (2.4.40), (2.4.42) and (2.4.43) respectively. Consider a further condition, that the motion takes place near a wall and consequently y <R. Then q~=x, q2=y, q3=y, h t = l , h2=l, hs=r, Vl = dxldt = Vx, V = Vy, V = 0. 2 3 The acceleration component from these data will be
(3.1.48) (3.1.49)
(3.1.50) (grad div V)1 = div Flax, where divergence of velocity is determined by the relation (2.4.44). So
The rotation of velocity component can be found with the help of (3.1.28): (rot 7 1 3 = av,/ax
a vxlay,
(rot
Vlz= O.
(3.1.52)
99
1 aVv avx '(rot rot V)I=. r L P ( a x ay)I From (3.1.51) and (3.1.53) we have
(AV)1=(grad div (rot rot
61
Similarly, the following expressions can be obtained for components along coordinate line q2:
w = av,lat+ 2
vx(av,/ax) + VY <av,>lav>;
Using the relations so obtained, the equations of motion can be written as:
avx  + v x  avx , ~ =  ~ . +v at ax
F+v,
ap v a div y 7 + a~ P axv (AV)l + Ti .  'l(3.1.59) ax a+~,av,=r. v, .a div f7 I ax a~ p ap+v(i,2++ a~ a~ ,J
where (A V)I and (A V)2 are given by formulas (3.1.54) and (3.1.58) respectively. This way we get different forms of equations of motion of a viscous fluid. This was done because experience in research shows that in some cases it is convenient to use one form of equation, in other cases another form. The analysis of particular types of fluid motion using the most suitable form of equation will be discussed in the following chapters. 2. Equation of Energy and Diffusion of Gas
2.1 Equation of diffusion Investigation of the motion of a dissociating viscous gas is related to the effect of diffusion of gas on this motion. This is due to the fact that diffusion is considered to be one of the basic equations of gas dynamics occurring in the study of an energy equation. The process of dispersion of concentration as a result of molecular transfer is called diffusion. This is a thermodynamically irreversible process which acts as one of the sources of loss of mechanical energy on the part of a moving gas medium. An equation of diffusion represents an equation of transfer of the ith
100
AERODYNAMICS
component of a gas mixture (this equation will be the equation of continuity for the same component). For simplification of research we assume that the difference between thermodiffusion and barodiffusion is infinitesimally small and the diffusion rate of the ith component in any direction n is determined by the equation (3.2.1) Qidn =  pDi (acilan), where ci is concentration of the ith component; Di is the coefficient of diffusion determined by the diffusion flow in the presence of a concentration gradient. For a mixture of gas components it is necessary to keep in mind the coefficients of binary diffusion corresponding to each pair of components, for example, for each pair of atoms and molecules of oxygen or nitrogen. The coefficient of binary diffusion in approximate calculations may be taken to be of a unique value 5 for each pair of components. Then Qidn=  pp (aci/an). Assuming that diffusion takes place along x, y and z directions, we have: Qidx= p fi(a~i/ax),Qidy= p ~  ( a ~ i / a Qidr=  p D(acilaz), (3.2.2) ~), or in vector form (3.2.3) Diffusion of a substance takes place in a region of low concentration. Consequently, acilan has a negative sign. Hence Qid will be positive. The equation of diffusion can be examined on the r assumption that motion is steady and takes place with respect to the cylindrical system of coordinates. Here the flow is supposed to be threedimensional and symmetrical about the x axis, i.e. the component of velocity V, =0. Let us separate out an elemental volume of 0 gas in the form of a ring with thickness dv and length dx (Fig. 3.2.1). Let the components of velocity be V, and Vr at a point P having coordinates x and P. Assume that the diffusion flow occurs in a radial direction only. Hence the flow rate of the ith component through Fig. 3.2.1. Elementary gas an internal surface of the element is m,= particle in axisymmetrical 22nrp V,s dx + Qid2~rdx,where ci and Qid threedimensional f o . lw are the concentration and diffusion flow rate of the ith component, calculated per unit area. The flow rate of the ith component through the external surface is
11 0
Ignoring the diffusion flow rate of the substance along the x axis, the gas flow rate through a lefthand element area perpendicular to this axis will be mx=p VXc 2ardr, j and through a righthand elemental area
Thus the flow rate of the ith component in a given volume will be equal to (pVxrci)/ax] 2ndrdx. As the quantity of gas in a given volume must not change the general inflow of the ith component will be equal to its consumption in chemical reactions. If the rate of transformation of the ith component in unit volume of gas due to chemical reactions is (Wchem)i [kg/m3.sec]the consumption rate of the ith component in a given elemental volume will be (Wchem)i 2zrdrdx. Then the balance of mass of the ith component in a given volume can be written in the following way:
[a
This equation is known as the equation of diffusion in cylindrical coordinates. Similarly the equation of diffusion for a plane flow can be obtained in the Cartesian system of coordinates x and y:
where Qid is found from (3.2.3). If a binary mixture of atoms and molecules is taken L'c~ cA+ CM = 1 and = ~ . hence Q A =  Q M ~ The value of (Wchem)i for a given reaction in a dissociating gas is determined by the formula of chemical kinetics (4.9.7').
2.2 Equation of energy
An equation of energy comes into the picture along with the equations of state, motion and continuity in the system of differential equations. The solution of these equations completely determines the motion of a gas. Consider a system of Cartesian coordinates and formulate the equation of energy for a fluid particle in the form of an elemental parallelepiped. This equation represents the law of conservation of energy according to which the change in total energy consisting of the kinetic and internal energy of particles in time dt is equal to the work done on an elemental particle by external forces plus the heat flow to it from outside. The kinetic energy of a particle having volume z=dxdydz is equal to (pV2/2)z and its internal energy is Upz (Uis the internal energy of unit mass of fluid). Consequently the change in total energy for the time dt will be
102
AERODYNAMICS
The work done by external mass (volume) forces during  motion of a  the particle in time dt can be expressed by the scalar product G . V, multiplied by the mass of particle pr and time dt. The vector of mass force is &Xi+ Yj+ Zk and hence (z.7) dt =(XV, + YV,+ZVz) pz dt. pz Let us now calculate the work done by surface forces. First we will consider the work done in time dt by the forces set up by stresses acting on the right and left sides. The work done by such forces on the left side is equal  to the scalar product g x V multiplied by area dydz and time dt. In a scalar product the vector of surface forces is
GX
=pxx i rxu j
+zxzk.
The work done by the surface forces acting on the right side is equal to
This expression has a positive sign because the forces on the right side are directed to the opposite side. Summation of the work done by all surface forces acting on the left and right sides gives:
Similarly the expressions for work done by the surface forces acting on the lower and upper faces and on the rear and front faces of the element can respectively be obtained by:
[a (, v ) / a y ] zdt g.
and
[a (&.@/az] zdt.
As the vectors of surface forces acting on the lower and rear faces are respectively equal to
103
The stresses in expressions (3.2.6), (3.2.6') and (3.2.6") are determined by the relations (3.I .5) and (3.1.16) respectively. The heat flow to a particle takes place through heat conduction, diffusion and radiation. Let qxdydz (where qx is the specific heat flow rate) represent heat flow due to heat conduction or diffusion to the partick through its left side in unit time. The amount of heat flow rate to the particle is q,dydzdt in time dt. Heat flow through the right side is equal to [qx+(aqx/a.~)dx] dydzdt. The amount of heat flow to the particle across both these sides is (aqx/ax) zdt. Similar expressions can be obtained for sides perpendicular to axes y and z. The resultant total heat flow to the particle will then be equal to
If heat transfer due to heat conduction is examined the heat flow rates, equal to the flow of heat along corresponding coordinate directions through unit area in unit time, can be expressed by Fourier's rule:
Qidx 6 qdy =
2
i
C
i
Qidz b;
Qid
2
i
104
AERODYNAMICS
and diffusion the heat due to radiation equal to ~ z d (E is the heat flow rate t due to radiation in unit time per unit volume) also flows to it. Equating the change in energy of this particle in time dt to the summation of work done by mass and surface forces and the heat flow due to heat conductivity, diffusion and radiation, we get the equation of energy:
c
i
ii div
(p
grad ci) + 8 .
During the motion of a chemically reacting mixture of gases the equation of energy must show the condition of heat balance, which includes heat appearing as a result of chemical reactions. However if we assume that the resultant enthalpy of gas mixture Z ~ iduring reactions remains unaltered, we i need not separately consider in equation (3.2.10) the liberation or absorption of heat due to chemical reactions. Substituting expressions for stresses from (3.1.5), (3.1.16) to (3.2.10) and ignoring the term for the work done by mass forces, we get, after suitable changes, an equation of energy for gas in the following form:
C
i
ii div ( p
i
J
I
Equation (3.2.1 1) shows the reason for change in the kinetic energy of a fluid. Besides heat conduction, diffusion and radiation, this energy varies due to the work done through compression div (p and work done by frictional force (terms containing the coefficientof dynamic viscosity p). The dissipation (from Latin dissipare"disperse") of energy is related to the losses of mechanical energy in overcoming frictional forces. Dissipation of energy occurs due to the fact that part of the mechanical energy will be unidirectionally transformed into heat. Therefore frictional forces are known as dissipating forces. The terms on the right side of equation (3.2.11) containing ~r form the dissipating function. For a twodimensional plane motion of a viscous medium the energy equation will be:
+ div ( A grad T) +
where
c
i
v=v,i+Vyj,
divV=aVxlax+aVy/a~,
grad T= (aT/ax) i (aTlay)j, grad Ci = (aci/ax)i+ (acilay)j. Let us transform the equation of energy (3.2.12). Multiply the first equation of (3.1.20) by V,, the second by Vy and add. Then
vx(aplax) + V ,
or
Two other terms on the right side of the equation can be transformed into the following form:
 div ( p F d i v 1) +
2 3
p (div 7))'.
Introducing a corresponding change in (3.2.13) and subtracting the equation so obtained from (3.2.12), we get d dp i p=
+div (A grad T )+
C
i
) (3.2.14)
where enthalpy i= U + ( p l p ) . In the absence of the heat transfer due to diffusion and radiation the equation of energy can be written in the form
At low speeds of gas flow, i.e. when the work done by frictional forces is small, the dissipating terms can be neglected. Moreover the work done by pressure forces (dp/dt%O) under such conditions is also small. In this case, in place of equation (3.2.15) we will have
(dT/dt)= (Alp c,) div (grad T ) . (3.2.16)
The quantity A/(pc,) =a, known as the coefficient of temperature flow, characterizes the intensity of molecular heat exchange.
Before writing down this system of equations we will study the equation of energy independently. According to the assumption of adiabatic flow the energy equation (3.2.14) can be transformed into di =dplp. (3.3.1) Keeping in mind that di=cpdT and taking into account the expressions cpcv= R and p = RpT, it can be shown that
p=Apk, (3.3.2) where A is a constant quantity typical for given conditions of gas flow. As we know, equation (3.3.2) is called an adiabatic (isentropic) equation. So in the present case the equation of energy coincides with an adiabatic equation. Then all the equations of the system can be written as follows:
Let us now study the system of equations for a more general case of motion of an inviscid gas at high speeds during which the specific heat changes with the temperature and the processes of dissociation and ionization may occur in the gas. Here, for the sake of generality, we acknowledge the possibility of radiation of energy by the heated gas. Then the thermodynamic process in the gas flow will no longer be adiabatic. Accordingly, the quantity E determining the radiating heat flow remains on the right side of the energy equation (3.2.14). Further, the equation of state should be taken in the form of (1.4.8) where the variation of average molecular weight pa"with temperature and pressure is considered. Assuming that the equations of motion and state do not change in form, the basic equations of the system will have the following form: dvx 1   . ap dt dvzdt dvY . a p . ] l = pax' dt pay' I 1 .dp+pdiv T=o; I p a z Y dt I Ro di dp p=pT, p=+E. Pav dt dt JI
a~
>
It can easily be seen that three more quantities, i.e, enthalpy i, average molecular weight of gas pavand heat flow due to radiation &, appear in this system in addition to the abovementioned six unknown quantities (V,, V V,, p, p, , 2"). Along with these quantities it is necessary to determine entropy S and also speed of sound a in studying the behavior of the gas flow. So the total number of unknown parameters figuring in the gas flow problem becomes five. Therefore it is necessary to add the same number of independent relations for these additional unknowns to the system of basic equations. These relations can be written in the form of general equations determining unknown quantities as functions of pressure and temperature:
Investigation of these functions is the subject of special departments of physics and thermodynamics. The solution of equations (3.3.4) to (3.3.9) determines the parameters of motion of an inviscid, dissociating and ionizing gas with consideration of radiation effects. Study of such motion is dealt with in the aerodynamics of radiating gas. Finally we will consider a still more general case of flow, taking in the effects of frictional forces and heat transfer. We assume here that chemical reactions take place in the gas. The main equations of this system will therefore be written in the following way (for brevity only a plane twodimensional flow will be considered):
%+pdiv dt
+ div (A grad T )+
C
i
ii
I I
This system is ,to be completed by the relations (3.3.5) to (3.3.9) and by the general relations for the coefficient of heat transfer (3.3.11) 2 =f6 ( P,T I ; coefficient of dynamic viscosity
/.1=f7 (PYT )
and specific heats
(3.3.12)
The last two quantities do not enter directly into equations (3.3.10). However, they are used in their solution so far as the determination of the thermodynamic characteristics of the gas is concerned. In addition, it is necessary to use the equation of diffusion (3.2.5) because the heat exchange due to diffusion is taken into account in the energy equation. At the same time it should be borne in mind that the concentration ci appearing in the equations of energy and diffusion is a function of pressure and temperature and may be expressed by the general relation The system of equations given above, consisting of the basic equations of gas dynamics and the corresponding number (depending on the number of unknown quantities to be found) of additional relations, is studied in the aerodynamics of viscous gas. This system, in principle, helps in the determination of the distribution of normal and tangential stresses and also the aerodynamic heat flows from a heated gas to the wall under interaction. The given system is simplified in some particular cases for which determination of the scheme of interaction processes is possible. This makes the solution of differential equations easy. In the process of solving these equations the necessity arises for additional relations for determination of the characteristic parameters of motion. These include, for example, relations for determining specific heats and the degree of dissociation as functions of pressure and temperature, formulas for calculation of frictional stress depending on velocity, etc. The solution of a system of gas dynamic equations, describing a flow over a given surface, must satisfy the given initial and boundary conditions of this flowinteraction. The initial conditions are determined by the values of the gas parameters for a given moment of time and obviously they have significance only for unsteady motion. The boundary conditions are imposed on the solution of the problem for both steady and unsteady flows and they must be fulfilled at every moment of time for this motion. The solution under this condition must be such that
1 10
AERODYNAMICS
the parameters so determined at the boundary dividing disturbed and undisturbed regions of flow are equal to the values of the undisturbed parameters. The second boundary condition is obtained by the nature of the gas flow on the surface being studied. If the gas is viscous and the surface is impermeable the flow is characterized by motion adhering to the surface (no separation). According to this condition of flow the component of velocity normal to the surface at any point over it is equal to zero and the resultant velocity vector coincides with the direction of the tangent to the surface at that point. It is known that the vector grad F [F( q ~q2, q3)=0 is the equation of the , surface in a flow, ql, q2, q3 being the generalized curvilinear coordinates] coincides with the direction of a normal to the surface. Then a smooth flow over the surface leads to the condition that the scalar product of this vector and the velocity~vector will be equal to zero. So the condition of smooth flow over a body can be expressed in mathematical form as
V .grad F= 0.
a~
a~ a~
i3.
a ~ .
wi3;
Fvx+aF,r+l.a_Fvy =,.
ax.
al
r ay
111
Other boundary conditions determined for any particular body can also be formulated. Here, boundary conditions for a viscous gas differ from those of an ideal medium. In particular, the solution of the corresponding equations in the case of motion of a viscous gas in the boundary layer must satisfy the conditions on the surface and on the external limit of the boundary layer. Experimentally it is found that the gas particles seem to adhere to the surface and hence the velocity over the surface is equal to zero. At the external limit of the boundary layer the velocity will be the same as in the free (inviscid) flow and the frictional stress will be equal to zero. 4. Integrals of Equations of Motion in an Ideal Fluid The differential equations written for the general case of motion cannot be integrated in their final form. But the integrals of these equations can at least be obtained for a particular case of ideal (inviscid) gas flow. The equation of motion of an ideal gas is written in vector form as:
E+grad
at
v2+rot  X   gradp. 1 Z. V V=
P
This equation can be obtained from the vector relation (3.1.22') in which the terms on the right side, accounting for the viscous effect, are supposed to be zero. The equation of motion in the form (3.4.1) was first obtained by the Russian scientist Prof. I.S. Gromyeko. In the presence of mass forces Gromyeko's equation takes the following form:
c=
Assuming the unsteady flow to be potential, we get rot v=0, grad p. Besides, we assume that the mass forces have potential U and therefore grad U, where
v=
grad U=au i+a uJ.+  auk. ax az If the medium is barotropic, characterizing a singlevalued relation between pressure and density (this takes place, for example, in the case of an adiabatic flow for which p= Apk) the ratio dp/p is equal to a differential for some func
a~
112
AERODYNAMICS
Taking this into account, equation (3.4.2) can be written in the following form: @/at) grad p +grad (V2/2)= grad U grad P. Replacing the derivative @/at) grad p by the quantity grad (&/at), we get (3.4.3) grad (ap/at) +grad (V2/2)= grad U grad P. Transferring this relation of gradients to the relation of corresponding scalar functions we get (ap~at) +(v2/2) +p+ U= where
c (t),
(3.4.4)
Equation (3.4.4) is called Lagrange's equation or Lagrange's integral. The right side of equation (3.4.4) represents the function of time independent of coordinates, i.e. it is unique for any point in a given potential flow. The terms on the right side have a simple physical meaning: V2/2 is kinetic energy per unit mass and P=Jdp/p is potential energy due to pressure per unit mass; Uis potential energy due to the position of fluid particle per unit mass. We recall the relation for potential function ap/as=Vs to understand the physical meaning of the first term, where Vs is the projection of a velocity vector along
S
the projection of inertial force due to the presence of local acceleration per unit mass and the product (aVs/at) ds is the work done by this force on segment ds, Then the derivative ap/at is equal to the work done by inertial force on the segment between points SOand s. It can be considered as the energy per unit mass due to the rate of change of velocity at a given point and due to the pressure in that condition. Then the expression on the left side of (3.4.4) represents energy per unit mass of gas. Thus Lagrange's equation establishes the fact that the total energy per unit mass of gas at a given time in potential flow is the quantity that is constant for all points in the flow. For an incompressible fluid, whose motion takes place under the action of pressure force and weight, the integral (3.4.4) is given by
This particular form of Lagrange's equation is known as Euler's equation. It represents the limitation of steady potential motion of a gas, indicating that the total energy per unit mass of gas has constant value at all points in the flow. In this way the constant in Euler's equation is not only unique for the whole region of flow but, unlike the function C(t) of Lagrange's integral, it is independent of time. Euler's equation for an incompressible fluid ( p = const) will be the same as (3.4.7)in form with the difference that the ratio p / p replaces Jdplp. Let us consider the more general case of a steady potential flow of gas. The equation of this type of motion will be given as grad (V2/2)+rot v x
v= grad P grad U

(3.4.8)
2fl
(3.4.8')
The right side of equation (3.4.8') is equal to zero if the vectors rot and V are parallel, i.e. under the condition that the vortex line and stream line coincide with each other. In that case
This equation was first obtained by I.S. Gromyeko. The constant CI will be the same over the whole region in which the condition of coincidence of the vortex line and stream line is satisfied. This flowinteraction is characterized by formation of vortices practically coinciding with the direction of stream lines near a wing. However, such regions are not always found in a flow. A flow is usually characterized by the presence of vortex lines and stream lines not coinciding with each other. In such condition the family of vortex lines is given by equation (2.6.1) and the family of stream lines is given by equation (2.1.6). The flow we are considering can be described by equation (3.4.8'). Let us take the vector of an arc in the form ds= dx .if dyajf dz .k belonging to a stream line or vortex line and determine the scalar product
d; grad [V2/2
(dplp)+ U ]=  di(rot V x V ) .

1 14
AERODYNAMICS
The left side of this equation shows the total differential of the three terms in square brackets. Consequently d [V2/2+ (dpip)+ U ] = dT(rot V x V). The vector product rot v x represents the vector perpendicular to the The scalar product of this vector and vector dFwill be vectors rot v a n d zero in two cases: when vector d i coincides with the direction of the stream line (trajectory) and when this vector coincides with the direction of vortex. In these two cases the equation
 
is valid, where C;!will vary depending on the type of stream line or vortex line. Integral (3.4.1 1) is known as Bernoulli's equation. It is obvious that the constant for different vortex lines, passing through points on given stream lines, will be the same as for the stream line. In the same way the constants will be the same for the family of stream lines (trajectories) and for the vortex through the points in which the stream lines pass. It is necessary to explain clearly the difference between these equations of Gromyeko's and Bernoulli's. Both of them are worked out for rotational (not potential) flow. However, the first equation shows the constant total energy per unit mass of gas in the whole region where vortex lines and stream lines coincide. The second equation establishes the limitation about the constant value of this energy along a given stream line or along a given vortex line. Then the constant in Gromyeko's equation will be the same for the whole region of flow. In Bernoulli's equation the constant is related only to a given stream line or a given vortex line. It is obvious that the two constants C I and C2 in a general case are not the same. From the above discussion the difference between these equations follows. At the same time the difference between Lagrange's and Euler's equations also becomes clear, namely that they are related to unsteady and steady vortexfree (potential) flow respectively. The difference between any two of the equations considered above is clear from the above discussion. Bernoulli's equation, related to the conditions along stream lines (trajectories), is very widely used in the study of the flow of a fluid or gas. It is known that the constant C2 of this equation (3.4.11) is determined for each of the stream lines being considered. If the steady flow is vortexfree (potential) Bernoulli's equation coincides with Euler's equation and therefore the constant will be the same for all stream lines, i.e. for the whole region of flow. Let us study some particular forms of Bernoulli's equation. For an incompressible flow under the condition that the function U = g y this equation can be written in the form
115
The effect of weight can be neglected in the case of a gas flow. Consequently we can assume U=O in equation (3.4.11) and in other integrals. In this particular case equation (3.4.12) is replaced by
Let us consider the motion of an ideal compressible gas. The heat transfer processes connected with the property of viscosity (heat conduction, diffusion) are absent in such a gas. Assuming that the gas does not radiate energy we will be dealing with the adiabatic (isentropic) motion of a gas. From equation of energy (3.2.14) it follows that equation (3.3.1) represents the motion of this inviscid gas without radiation (8 =O). Hence Bernoulli's integral is
( V2/2)f i= C.
(3.4.14)
Bernoulli's integral in this form represents the equation of energy for isentropic flow. According to this equation the sum of kinetic energy and enthalpy of a gas particle has constant value. Assuming that i= cPT=c, p/(pR), C cV R and k =c,/c, we get , =
Hence
Bernoulli's equation for an ideal compressible gas happens to be the theoretical basis for investigation of the limitations of isentropic flows of a gaseous medium. This is examined in Section 6.
5. Aerodynamic Similarity
5.1 Concept of similarity The aerodynamic characteristics of flight vehicles or of their individual elements can be determined theoretically as well as with the help of experimental research. Theoretical methods are based on the application of a system of gas dynamic equations which are solved with respect to a flow over a body of a given form having in general arbitrary values of absolute sizes. In running experiments to obtain the aerodynamic parameters, which can be directly used for further research and for confirmation of theoretical findings, the actual body cannot always be used due to its large size. It is often
1 16
AERODYNAMICS
necessary to use a model of the body on a reduced scale. There the problem arises of transferring the experimental findings to the real body. Dimensional analysis and the theory of similarity take care of this problem. They establish the conditions that must be observed in experimental work with models and indicate the characteristic, convenient parameters determining basic effects and flow regimes. Let us assume that the drag force, measured on a model in an aerodynamic wind tunnel, is equal to X = cXmqmSm , according to equation (1.3.5). We will now determine how the result so obtained can be used to estimate the drag force on the prototype body which is given by the formula Xp=cxpqpSp where cXpis the unknown coefficient of drag for the actual body and qp and Spare the given velocity head and characteristic area respectively. Dividing Xp by X,, we get
It can be seen from this expression that estimation of the force on the real body Xp from the experimental value of Xmcan be achieved only on the basis of equality of aerodynamic coefficients CX, and cxpbecause the values of qmSm and qpSpare uniquely determined by the given values of velocity heads and characteristic areas. In this case the flows over model and prototype will have the property of aerodynamic similarity which, in this case, yields force charaoteristics of one flow (force Xp) from the given force characteristics of the other flow (force Xm) by simple calculation involving transfer of one system of measurable quantities to the other. The requirements that bring about the equality cnm cxp in a given case, = and in the general case for other aerodynamic coefficients, are established by the theory of similarity and dimensional analysis. They follow either from the physical nature'of the problem being investigated or from the corresponding differential equations of aerodynamics. Examining the expression for aerodynamic coefficient
obtained from (1.3.2), it may be noted that this coefficient depends on nondimensional geometric parameters and on nondimensional quantities like the coefficients of pressure and local friction. It follows from this that the aerodynamic coefficients for the real body and for an experimental model of different absolute sizes are maintained constant if these bodies are geometrically similar and the distribution of coefficientspand cfx is kept the same over their surfaces.
117
Consider a steady uniform flow of compressible gas over a body without heat transfer. It follows from physical considerations that the coefficients p and cfx for the given form of body and known values of angles of attack and banking will be functions of velocity V pressure p,, density p,, coefficients , of dynamic viscosity lr,, specific heats of gas c,, and c,, in an undisturbed flow and some linear characteristic dimension of body L. Hence the coefficient of drag force will also depend on these parameters, for which the functional relation can be written in the form cx=f (V,, p,, p,,
P9 
CP,,
cv,, L).
Since this coefficient is a nondimensional quantity the parameters of this function in the above expression must be nondimensional. From the general discussion on dimensional analysis it follows that the seven dimensional variables of function c, can be reduced to three parameters which represent nondimensional combinations of V p,, p,, /roo, cp,, CV,, L. This is because there , are four independent measurable quantities; mass, length, time and temperature. These nondimensional combinations can be expressed in the following form: ~ , / d k , p,/p, = V,/a, =M,, Mach number of an undisturbed flow; Vp L/,u,=Re,, ,, Reynold's number calculated on the basis of undisturbed flow parameters and the characteristic linear dimension L;cpw/cv,=k,, index of adiabatic process. In the expression f ~M, it is assumed that dk, p,/p, =a,, the speed of r sound in an undisturbed flow. According to the general expression for speed of sound a2=dp/dp. Assuming the adiabatic nature of the propagation of sonic disturbances in a gas so that p=Apk (where A is a constant to be determined by initial conditions), we have
For the propagation of sound in an undisturbed flow it will be given by a%=km (pmlpoo). In this way the ratio=, ,v / V,/a,. All the other nondimensiona1 combinations except M,, Re,, k, composed of the given seven parameters or, in general, of any other quantities determining them will be functions of combinations of M,, Re, and k,. The coefficient of drag can therefore be written as
cx =f (Mm,Re,, k,).
(3.5.3)
Similar expressions can be formulated for other aerodynamic coefficients. It follows from these expressions that in equating M,, Re, and the parameter k, of the model with those of the prototype the aerodynamic coefficients of geometrically similar bodies are taken as identical. So we arrive at the important conclusion of the theory of similarity and dimensional analysis according
118 AERODYNAMICS
to wliich the necessary and sufficient condition of aerodynamic similarity is identical values of the nondim'ensional parameters forming the socalled basis of analysis, i.e. the system of nondimensional quantities determining all remaining flow parameters. These nondimensional combinations are known as the similarity parameters. The similarity criteria given above have a particular physical meaning. The expression a2=dp/dp shows that the speed of sound can be taken as a parameter depending on the compressibility of a gas, i.e. depending on the property of gas whereby it changes its density with any change in pressure. Therefore the Mach number happens to be that nondimensional similarity criterion that characterizes the relative value of the compressibility effect on a gas flow. Reynold's number represents by itself a criterion to account for the relative value of viscous force on a moving gas and the parameter k= , c,,/c, determines the peculiarities of flow that are based on the thermodynamic properties of a gas.
5.2 Similarity criteria to account for viscosity and heat conductivity effects In more general cases of flowinteraction, characterized by the effe~t a of series of other physical parameters on the aerodynamic and thermodynamic properties of the flight vehicle, the criteria of dynamic similarity are more complicated and of various types. The other method of the theory of similarity and dimensional analysis can be applied to establish these criteria. This method is based on the use of the equation of motion of a viscous heatconducting gas. Let us write down these equations in nondimensional form, i.e. in a form such that the parameters entering into the equations (velocity, pressure, temperature, etc.) are expressed in terms of some characteristic parameters. For a given flow these parameters are constant quantities and they represent the characteristic quantities of the flow. In the name of characteristic quantities we select the flow parameters of undisturbed flow: velocity V,, pressure p,, , , density p,, temperature T coefficient of dynamic viscosity ~ l(or corresponding value of v,) etc. It is necessary to bear in mind that only two of the three parameters p,, p,, T, may be given and that the third can be calculated from the equation of state. The characteristic time indicating a period of unsteady flow will be quantity t,, the characteristic length will be the characteristic linear size L (for example, length of the body in a flow). Gravitational acceleration g can be taken as the characteristic acceleration of mass forces. The nondimensional parameters of length and time may be given by
and the nondimensional coefficients of speed, pressure, density, viscosity and mass forces can be written in the following form:
1 19
Let us introduce nondimensional quantities in the equation of motion (3.1.17)and the continuity equation (2.4.2). In this process of transformation we take the first equation of the system (3.1 17) because the remaining two equations can be written similarly. The abovementioned equations in nondimensional form will have the form:
ai
ax
aY
%)
az
From the characteristic quantities entering into these equations it is possible to form a series of nondimensional numbers characterizing the similarity of gas flows. These numbers are known by the names of the scientists who first conceived them and are written in the following form:
Sh = V , t,/L, Strouhal number; 7 Fr = V ~ I ~ LFroude number; , M = V,/a,, Mach number; (3.5.6) I Re =(V, L p,/p,) = V,L/v,, Reynolds number 1 (index "co" for M and Re is dropped). J Here a,= d k , ~ ~ , speed of sound in an undisturbed flow (k, and T, is the are respectively the adiabatic index and gas temperature in an undisturbed flow). Introducing these numbers in the equations of motion and continuity, we get:
=!
120 AERODYNAMICS
Let us now represent the equation of energy (3.2.14) in nondimensional form, neglecting the terms for radiation and diffusion. The following nondimensional parameters
(3.5.9) T= TIT,, cP= cPlc,,, A =A/A,, will be introduced, where c, and A, are respectively the specific heat and coefficient of heat conductivity of the gas in an undisturbed flow. Recalling that di=cpdT and expanding the total derivative dT/dt, after necessary substitutions we get
 U (div R' +4 3 ,
;) :
which helps us in comparing the relative values of interaction of viscosity and heat conductivity or, in other words, in establishing the relation between heat flow due to friction and molecular heat transfer. Then, taking into account that p,/p,= RT,= c (1  Ilk,) T,, one can write ,
p C,
Now we perform the transformation of some additional relations from the system of gas dynamic equations to the nondimensional form (see Section 3):
+'+

div (2grad
3.
(3.5.10)
121
Let us suppose that two flows over two geometrically similar surfaces are being investigated. For such surfaces the nondimensional coordinates of respective points are the same. This is actually the necessary condition of aerodynamic similarity of flows. To satisfy the sufficient condition of similarity the equality of nondimensional quantities of gas dynamic parameters (velocity, pressure, density, etc.) must be observed at similar points in the two flows. Since the nondimensional parameters occur simultaneously as the solutions of the system of equations (3.5.7), (3.5.8), (3.5.101), (3.5.12) to (3.5.16) it is obvious that the abovementioned equality will be observed on condition that the systems of nondimensional equations and the nondimensional initial and boundary conditions for each of the flows be identical. Examination of the systems of nondimensional equations for the two flows shows that these systems will be identical if: 1) similarity parameters are equal: Frl =Frz, Re1 =Rez, MI =M2, Shl =Sh2; 2) the equality of Prandtl numbers Prl = Pr2 is fulfilled, i.e. (cpW m ~ l m j ~ , U ~ / ~ ~ ) Z , p =(cpm (3.5.18) (3.5.17)
and the equality of adiabatic indices for gas in the two flows is satisfied:
3) each of the equations (3.5.13) through   determines the depen(3.5.16) deuce of the nondimensional quantities iav,A, p or cp on the nondimensional quantities p a n d Tand on the parameters in (3.5.17) through (3.5.19). However, in the case of a dissociating gas such relations do not hold because the nondimensional criteria of the type (3.5.17) through (3.5.19) or of any other form cannot be separated. Therefore the corresponding nondimensional equations (3.5.13) through (3.5.16) for the real flow and for the flow over the model will not be identical and complete dynamic similarity cannot be obtained. Two particular cases may be mentioned where this similarity is satisfied. The first is the case of a nondissociating gas flow for which the average molecular weight remains constant (pav,=,uavJ and the heat content, coefficient of heat conductivity and viscosity change with the temperature according to a law of the  y ~ e =aTX. In this case equations (3.5.14) through (3.5.16) for t y quantities A, ,u and are replaced by the corresponding relations of nondimensional temperature %TIT,. The second case is connected with a gas flow at comparatively low speeds when parameters 1 p and c, do not depend , on temperature. The corresponding values of these parameters will be identi
cp
122
AERODYNAMICS
cal for the actual flow and the flow over the model. In this case the system of equations consists of the NavierStokes nondimensional equationsof continuity and energy and also an equation of state. The boundary conditions imposed for the solution of nondimensional equations may introduce some additional similarity criteria. This is not connected with the condition of unseparated flow, which does not carry any new similarity criteria. In reality this boundary condition in nondimensional terms has the form
and will be one and the same for both the actual flow and the flow over the model as a consequence of the geometrical similarity of the surface in the flow. However, the boundary condition for temperature, indicating that the solution for temperature must satisfy the condition that T=T, (Tsu is the temperature of the surface), poses additional criteria for similarity. In practice it follows from the boundary conditions for actual and model surfaces with the corresponding forms TI=(Tsu)l and T2= (Tsu)2that the nondimensional temperatures % = (T,)I and %= (zu)&By the similarity condition we have = T2 and hence the equality (T,)I = (Tsu)z must be observed. Thus the boundary condition for wall temperature leads to an additional similarity criterion ( T ~ ~ I= (Tsu/Ta)z. T~)~ (3.5.20) The nondimensional gas dynamic parameters on the surface of a body in a flow, as seen from the system of nondimensional equations [under the condi tion that equations (3.5.13)(3.5.16) determine h,A, , and c, as functions of u TI,depend on the nondimensional coordinates and time and on the similarity criteria (3.5.17)(3.5.19). In particular, nondimensional pressure can be expressed as a function
(PIP,)
%,x, Y , 2, t).
 
(3.5.21)
At a given moment of time the nondimensional coefficient of drag force can be determined from the known distribution of pressure ex= X/q, S = q2 (Fr, Re, M, Sh, Pr, k,,
G).
(3.5.22)
This expression, being more complete than (3.5.3), determines the relation of the aerodynamic drag coefficient to the nondimensional similarity criteria. However, the relation (3.5.22) does not reflect all the peculiarities of a dissociating gas flow because it is derived from the simplified equations (3.5.13) to (3.5.16). Therefore formula (3.5.22) is less accurate for such a flow than for a nondissociating gas flow but it at least represents partial similarity. Similarity criteria, which determine the nondimensional aerodynamic coefficient, have a particular physical meaning and characterize the real factors affecting aerodynamic forces.
123
The Froude number happens to be the similarity criterion accounting for the effect of the weight of a fluid on its resistance. From the equation of motion in nondimensional form it can be seen that Fr is equal to the ratio of quantity V~/L,based on the effect of inertial forces, to the characteristic mass force g. Equality of Froude numbers for the real body and a geometrically similar model indicates that they will have identical drag coefficients, which are based on the effect of the weight of a fluid. This similarity criterion does not have any significant value in the study of gas flows because the effect of the weight of a gas on motion is negligibly small. However, this criterion may have considerable value in hydrodynamics, in particular in experimental research on the wave drag of different types of ships. For the motion of bodies in a real fluid aerodynamic forces depend on viscosity. The viscous force is characterized by the Reynolds number, which can be obtained as the ratio of quantity VL/L, which takes into account the effectof inertial forces, to the quantity V, Vw/L2,accounting for viscous effect. If the Reynolds number for two geometrically similar flows is kept equal the condition of partial aerodynamic similarity accounting for viscous effects is fulfilled. The coefficients of frictional drag for the real body and the model will be equal under this condition. The similarity criterion based on the Mach number is obtained from the ratio of quantity V&/L to the parameterp,/p,L, which takes care of the effect of pressure forces on the compressibility of gas. The partial similarity of two flows of compressible gas over geometrically similar bodies will be observed if the Mach number for both flows is identical. In the study of the problems of an unsteady flow the Strouhal number has an important role. This number is obtained by comparing inertial forces to the forces due to unsteady effect, i.e. it is obtained from the ratio of quantities V&/L and V,/t,. Two unsteady flows over the actual body and the model will have partial aerodynamic similarity if the Strouhal numbers are identical for both. Similarity criteria based on the Prandtl number and the ratio of specific heats are determined for given requirements of the physical properties of gases in the real flow and in the flow past the model. The gases used for these two flows may be different but their physical characteristics must be such that the equalities Prl =Pr2 and k , ~ =km2 are satisfied. The Prandtl number depends on the coefficients of dynamic viscosity and heat conductivity. The coefficient of viscosity represents gas properties related to molecular momentum transfer and the coefficient of heat conductivity characterizes the intensity of molecular heat transfer. Prandtl's criterion Pr = pcc,cp,/A, determines the mode of transfer of the energy of molecular exchange to heat energy. For gas, we have Pr < 1. The nondimensional coefficient of aerodynamic force or heat transfer is the complicated function of a series of similarity criteria, each of which reflects
the effect of some given physical process. Complete similarity between the real flow and the flow over the model can be satisfied only if all the similarity criteria are fulfilled. In practice, however, this cannot be observed because some of these criteria are of incompatible nature. Let us take, for example, the Reynolds number, Froude number and Mach number. To satisfy similarity of frictional forces it is necessary that VlLl/v~= V2Lzlvz. If the coefficients are assumed equal, i.e. v l = v 2 for flows over the real body and the model the velocity of flow over model V2= vl(L~/Lz),i.e. the velocity of flow over the prototype is higher than the velocity of flow over the model in the same ratio as the size of the prototype to the model. To satisfy the similarity of forces based on gravitational weight it is necessary to observe equality of Froude numbers, i.e. Vf/(L~gl)= Vi/(L2g2), from which it follows that if the experiments are conducted at identical g, velocity of model flow V2= V I ~LX. the velocity V for the scaleddown Obviously 2 model must be appropriately lower, not higher than Vl. Finally, in observing equality of Mach numbers we will have V l / a ~ = V2/a2. Assuming, for simplicity, that az=al, we get the condition of equality of velocities in the flows over the model and the prototype. It is obvious that all these criteria cannot be satisfied at one time and hence only partial similarity can be observed. It should, however, be noted that there is no necessity in practice to satisfy all similarity criteria because their effects in some or other specific case of motion are not identical. For example, the effect of frictional forces and pKessure on the interaction of body and gas flow will be more appropriate than weight forces. Therefore the criteria of Re and M have more significance in such cases than Fr. That is why the Froude number is not considered as a similarity criterion in such cases. If, at the same time, the velocity of motion is not high the effect of pressure forces related to the compressibility of gas is negligibly small. Then it is possible to ignore the similarity criterion of the Mach number and assume that the aerodynamic coefficients depend on the Reynolds number. Aerodynamic force, moment and heat flow from gas to the surface are the result of interaction between the moving gas and the body where different processes like friction, compression (or expansion), heating, variation in physical properties, etc. take place simultaneously. Therefore it is necessary to try to satisfy the maximum number of important similarity criteria. For example, it is desirable that equalities of the Reynolds numbers and Mach numbers as between the actual flow and the model flow be fulfilled simultaneously, i.e. Re1 =Re2 and M I =M2. This is particularly important during the investigation of aerodynamic forces for bodies with large surfaces, which could be made up of various types of components depending on friction and normal pressure related to compressibility. The fulfillment of the above conditions can be observed in running experiments in aerodynamic wind
125
tunnels of variable density. If the experiments are carried out in a gas flow in which the speed of sound is the same as in the actual flow (az=al) the condition of equality of Mach numbers results in V2= Vl. Then, using equality Re1 =Re2 or, in other words,
we get the condition of L2p~/,u2=Llpl/,u1.Since ,u2=,ul it can be established that the gas density of flow in an aerodynamic wind tunnel should be p2= pl(L11L2). Assuming that temperatures in the actual and model flows are equal (Tz=Tl) we get, using the equation of state, the condition of pz= pl(Ll/Lz). Hence for simultaneous fulfillment of similarity in frictional forces and pressure forces accounting for compressibility, i.e. to satisfy equalities Re1 =Re2 and MI=M2, it is necessary that the static pressure in the gas flow in the wind tunnel exceeds the pressure in the actual flow by the same ratio as the size of the prototype to the size of the model. Wind tunnel design allows us to regulate static pressure in the model gas flow in a certain range depending on the size of the model in the flow. The effect of heat transfer may be neglected as an approximation in certain known cases of determination of forceinteraction. Here the aerodynamic coefficients will depend on Re, M and Sh. If the experiments are conducted in a gaseous medium for which koo2=km1, then
c,=
f (Re, M, Sh).
(3.5.23) (3.5.24)
In this.section we will examine onedimensional steady isentropic flows of compressible gas. The study of such flows is of great practical significance because it allows us to establish the connection between the jet cross section and variations in gas parameters and at the same time helps us to find how to control the flow by changing the form of the jet (or canal). We will also examine the wide range of gas dynamic relations consisting of the basic mathematics for calculation of isentropic gas flows with constant heat coefficients.
6.1 Form of gas jet Let us consider a steady flow of ideal (inviscid) gas in a jet with low expansion and small curvature. The flow in this jet can be taken as onedimensional, characterizing the variation of parameters depending on one linear coordinate of a point along the jet axis. For a steady flow the parameters determining this flow at every point will be constant at any moment of time. If the width of the jet is small compared to the radius of curvature of the
axial line the pressure gradient across the section can be neglected and the pressure at any two points of the jet cross section can be considered equal. Study of this onedimensional steady flow of compressible gas leads to the most simple approximate solution of gas dynamic equations. Along the jet axis the condition (2.4.51) of constant mass flow is satisfied, i.e. ~ ~ V I S I = p2V2Sz=p3 V3S3 = ..., or p VS= const, where indices "l", "2", "3" correspond to the gas parameters over the control surfaces in terms of cross sectional areas SI, S2, S 3 . . . .Taking the logarithm of both sides, we get In p +ln V +In S=const. Differentiation of this expression gives
where
Recalling the formula for the speed of sound a = Jdp/dp and using relation (3.6.3), we transform (3.6.1) into the form
where M = V/a is the Mach number at a given cross section of the jet. Let us assume that speed increases along the jet (dV>O) but that it remains subsonic and hence M < 1. From (3.6.4) it can be seen that for this case the derivative dS/dV<O. This shows that downstream of the flow the jet diminishes in cross section. Conversely, for a subsonic flow of decreasing velocity ( M e 1, dV<O) the cross section of the jet increases, which is indicated by the inequality dS >O as seen from (3.6.4). Let us now examine a supersonic flow (M > 1). If the velocity diminishes along the jet axis it can be seen from (3.6.4) that differential dS< 0 and hence the jet decreases in cross sectional area. In the opposite case of increasingjet velocity the value of dS>O, i.e. the jet expands. Consider a nozzle of diminishing cross section, opening out again at the tail end: a subsonic flow, under given conditions, will be accelerated in the initial section and reach sonic velocity at the narrowest point [here, dS=O and hence from (3.6.4) we have M= 11. Thereafter it will be supersonic. Nozzles in rocket engines, gas turbines and aerodynamic wind tunnels are designed in this way to obtain supersonic flows.
Let us consiiler a gas jet passing over some surface (Fig. 3.6.1). The para, meters for the undisturbed flow region in a jet are denoted by V p,, p,, T i, a,, etc. and those for the disturbed flow region by V, p, p, T, i, a, , etc.
Equation (3.4.14) can be used to find the velocity at any arbitrary section of the jet and constant C can be determined from the condition of the undisturbed flow parameters: C= (V2/2) Ii,. (3.6.5) Then we have (V2/2)+ i= 2,2)+i (V I from which
cay
'
Thus constant C, in the physical sense, can be regarded as the stagnation enthalpy. Taking this value of C, the velocity in the jet will be V= 4 2 (io  i). (3.6.8) Enthalpy io corresponds to pressure po and density po for stagnation conditions and can be determined from the condition
v : i ~ = PO .v'+i,=+.~ 2~ = 2
Expressing io in terms of po and po, we get V2 + 2
k
kl
p,
p,'
hence
Keeping in view that the speed of sound ao .\rkPo/po, for the condition of complete stagnation we find
From (3.6.8) it follows that velocity along the jet increases as enthalpy decreases and hence more heat is converted into kinetic energy. Maximum velocity is achieved when enthalpy i=O, i.e. when the whole of the heat is spent on accelerating the gas flow. The value of this velocity is
Using the concept of maximum velocity, the relation of velocity at any arbitrary section can be written in the following way:
V = Vdl
 (ilia) ,
(3.6.16)
At the narrowest critical section of the jet a speed equal to the speed of sound is achieved. This speed is called the critical speed and is denoted by a*. This critical speed corresponds to the critical pressurep* and density p*. From Bernoulli's equation it follows for the critical section that
129
Equation (3.6.10) may be used to determine the local velocity of speed of sound a. ~ n t r o d u c i n ~ it the quantities a2=kp/p and a,2=kpolpo, we get into
Let us introduce notation 1 = V/a* for relative velocity. Dividing equation (3.6.21) by V 2and keeping in mind the ratio V / a = M , we get the relation between 1 and M:
1 2=
From this it follows that at that section of the jet where velocity V is reached the Mach number M = co. The corresponding value of A = Amax can be found from (3.6.23) under the condition of M+co
It is obvious that at the critical section, where M = 1 , we have A= 1. At any arbitrary section characterized by the value 1 < M < co the relative speed
6.3 Pressure, density and temperature It follows from (3.6.17) that pressure at any arbitrary section of the jet will be
and therefore
130
AERODYNAMICS
Consequently we get
P=P"
=pm n ( M ) . 71. ( M w )
(3.6.30)
From the adiabatic equation p/pk=p,/p,k in which p is replaced by formulas (3.6.26) and (3.6.28) the relation for density will be
p=po
(1
I/@1)
I/(kl)
= po(l + k $ ~ 2 )
=pee ( M ) , (3.6.31)
where function E of Mach number M can be determined by density ratio plpo. Using the equation of state
(3.6.32)
and the relations for pressure (3.6.26), (3.6.28) and for density (3.6.31), we get the formula for temperature:
k
M)' =To 7 ( M ) ,
(3.6.33)
where function z ( M ) is determined from the temperature ratio TITO. The tables of gas dynamic functions n OM), E (M) and z(M) for various values of index k from 1.1 through 1.67 are shown in [53]. For density po and temperature To the corresponding relations can be obtained with the help of equations (3.6.31) and (3.6.33). Applying the conditions of undisturbed flow to these relations, we get
At the critical section of the jet M = 1 and hence, from (3.6.28), (3.6.31) and (3.6.33) we get the following formulas for critical valnes of pressure p*, density p* and temperature T*:
131
where n (I), E (I), z (1) are values of gas dynamic functions at M= 1. From the above formulas applicable for any speeds approximate relations can be obtained for cases where Mach numbers M are very high. From (3.6.30) it follows that at M, B 1 and MB 1 it can be written that
we get the following approximate formula of local velocity under the conditions of M  9 1 and M$ 1:
6.4 Gas flow from reservoir Formula (3.6.12) helps to determine the velocity of a gas flow through a nozzle from a reservoir (Fig. 3.6.2), where the flow parameters are determined by the stagnation conditions corresponding to the speed of gas flow in reservoir VzO. Such conditions are practically satisfied if the critical cross section of the nozzle is fairly small compared to the cross section of the reservoir. At the critical section of the nozzle, where the value of derivative dS/dV=O, the Mach number M, as seen from (3.6.4), is equal to unity, i.e. the speed at this section is equal to the local speed of sound:
This velocity relation is satisfied under the condition where the pressure in the reservoir, according to (3.6.36), is not below the value
If the nozzle opens to atmospheric conditions with pressure pa= 1 atmosphere (this pressure is called adverse pressure) motion of gas along the nozzle is possible under the condition that p* 2 1 atmosphere. With p* = 1 atmosphere and lc= 1.4 the pressure necessary at the reservoir to achieve sonic speed at the end of a converging nozzle will be
132
AERODYNAMICS
x 1.9 atm.
After achieving critical parameters in the narrowest section of the nozzle the further reduction of adverse pressure (p,<p*) has no effect on the values of a* andp* and these values depend on reservoir conditions only. However, conditions for occurrence of a supersonic speed of gas flow in the divergent part of the nozzle may arise during this process. This regime of isentropic flow is
Medium
pa
Pa
To
discontinuity
Y
characterized by the curves 1 in Fig. 3.6.2, indicating variation of ratio p/po and Mach number M along the nozzle. This regime of flow is found where adverse pressurep, is equal to pressure pi') or less than this at the exit of the divergent part of the nozzle. Then changes in the adverse atmospheric pressure have no effect on gas parameters at the exit of the nozzle. Let us look at one more important regime of isentropic flow along a nozzle. Assume that the adverse pressure increases to the value pi2)>pL1) at which pressure at exit pi2) reaches the same value as pi2). The subsonic flow so formed can be characterized by lines 2 shown in Fig. 3.6.2. The relative velocity L (M) of this section can be found from equation (3.6.23) by selecting one of the two solutions having value L < 1. A further increase in adverse atmospheric pressure to pL3)creates conditions where the pressure at the exit will be the same as atmospheric At the throat section the pressure increases beyond critical value p*; the
corresponding velocity at this section will be subsonic ( V c a*). Therefore the speed in the diveigent part of the nozzle will diminish and remain subsonic (curve 3 in Fig. 3.6.2). For values of pressure at exit less thanpA2) and more thanpL1) the flow does not remain isentropic. Discontinuity of flow appears at some section of the nozzle, associated with sudden stagnation of the flow across it. Because this stagnation has an irreversible adiabatic character the supersonic flow changes to a subsonic flow through shock. Variations in pressure and Mach number M along the nozzle in this region of flow are characterized by curves 4 in Fig. 3.6.2. The separation surface (shock wave) must lie at the section with the Mach number that could have given pressure p, behind this surface, providing pressure pi4)(PA') <pL4)<pA2))at the exit as a result of further expansion of the gas. Thus analysis of the gas flow in this case must be carried out in light of shock wave theory. The basic concepts of this theory will be discussed in Chapter 4. At any arbitrary section of the nozzle including its exit section the supersonic speed can be determined by formula (3.6.12). For a given Mach number Me =M, and pressure pe =p, at the exit cross section the required reservoir pressure po can be determined by (3.6.29). Let us write the equation of mass flow
from which we find that the ratio of specific mass flow q =pV through section S to the specific mass flow q* =p*a* through the throat section S* is
After determining the ratio of po/p* from (3.6.37) we get relation (3.6.45) in the following form:
The function (A) is called the mass flow density. According to formula (3.6.23) it is possible to determine this function in terms of Mach number M:
The tabulated values of function q in the range of k from 1.1 through 1.67 are shown in tbe work [53]. From these relations it follows that number
134
AERODYNAMICS
A (or
M) at some section S of the nozzle happens to be a function of the ratio of areas S*/S and does not depend only on the gas parameters in the reservoir. The variation of the specific flow7 (or ratio of areas S*/S) as a function of A is shown in Fig. 3.6.3. At a given critical section of the nozzle in the subsonic region of ii flow (A< 1) the increase in velocity is achieved by reduction of area S. On the other hand, in the supersonic region it can be achieved by A reduction of the nozzle section area S. There05 . fore a reduction in reservoir pressure does not affect the value of A or M at the nozzle exit. In the case being considered, as can be seen 0 2.0 from (3.6.29), the pressure at exit p e = p , will K 1 change in accordance with the variation in Fig. 3.6.3. Variation of specific pressure po. flow rate for gas flow. The weight flow rate of gas from the reservoir can be determined by the formula
Determining values of p* and a* by the formulas (3.6.37) and (3.6.18) respectively, we get
This kinematic equation (3.6.48) contains all the information on the onedimensional motion of an incompressible fluid. According t o this equation flow velocity V is inversely proportional to the area of the cross section of the canal. The pressure for a Bernoulli's equation steady flow case can be obtained from (3.4.12) or (3.4.13).
6.6 Dissociating gas flow from reservoir (onedimensional problem) Let us consider the flow calculations for a nozzle connected to a reservoir of gas at a high enough temperature for balanced dissociation to take place. We will assume that the distribution of velocities and pressure at every cross
135
section of the nozzle is uniform and depends only on the location of the section, i.e. the flow is considered to be onedimensional. We will also assume that there is no frictional loss in flow and no shock wave, i.e. the flow is isentropic. The given parameters usually consist of the parameters for the stagnation condition: pressure, temperature in the reservoir (the density, enthalpy or entropy are usually calculated from this), velocity (Mach number) at the exit section of the jet and the area of this section. The aim of this flow analysis is to determine the gas parameters at each section of the nozzle and the corresponding cross sectional area from the above data. In tackling this task it is necessary to account for the thermodynamic relations of flow parameters, not only with temperature but also with pressure. To do this the diagram of "enthalpyentropy" developed for the condition of balanced dissociation or the relevant thermodynamic tables can be used. Equation (3.6.7) complements the tables and the diagram inasmuch as it helps us to determine the local velocity of isentropic flow from a known value of enthalpy and vice versa. The following sequence may be observed in this calculation: Enthalpy io and entropy So are obtained with the help of the iS diagram (see Fig. 1.4.6) or from tables of thermodynamic functions, using known values of po and To at the reservoir. For a given velocity VA at the exit the corresponding value of enthalpy can be obtained from the equation io= i, + (V2/2)in the following form:
i~ = io  (V2,/2). (3.6.49) Then we assign a series of values to enthalpy i~ (iA<i ~ iO) and, based on these < values, we find velocities
vK=d2(ioix;). (3.6.50)
s *Is
,DO
lo'
As the flow is isentropic and consequently So=const for all sections the corresponding gas parameters (pressure PK, temperature TK, density p~ and specific mass flow rate pKVK) can be obtained for different values of iK from the i S diagram using the above value of So. To determine the flow parameters at the critical section it is necessary to draw the curves of variation of velocity V ~ = f i i ~ from equation (3.650) ( ) and the curves of variation in the speed of sound a ~ = f 2( i ~ from the ai diagram ) (see Fig. 1.4.8). The point of intersection
2 4 M Fig. 3.6.4. in area of cross sKtions of nozzle during flow of dissociating air:
 dissociatingair;
   perfect
gas at k= 1.4.
of these curves gives the values V*=a* and i*. After determining density p* from i* and SO the specific flow rate through the critical section can be calculated. The area of the critical section can be found by the formula
Mach numbers at corresponding sections are determined by the relation MK=V ~ l a ~ . T/& ;)/Po; p/po The results of this calculation show 10" that with an increase in Mach number the temperature along the nozzle diminishes at a lower rate than for an undissociated gas. This is explained by the fact that with the increase in velocity and consequent cooling liberation 1 of energy of recombination, making for additional heating, takes place in the dissociating gas flow. This rise in temperature leads to a reduction in density, which has an effect on the reduction in pressure by comparison with the pressure in an undissociating gas. So in the presence of dissociation a given velocity (Mach number) than without dissociation. The cross secdissociating air;    perfect air at tional areas of the nozzle are also k=1.4. found to be bigger with a dissociating gas flow. The results obtained for a particular case of flow from the reservoir with pressure PO= 100 kg/cm2 and temperature TO 12,000"K are shown on Figs. = 3.6.4 and 3.6.5.
Fig. 3.6.5. Variation in gas parameters in flow through nozzle:
'041333
1050
1. Physical Process of Appearance of Shock Waves The distinguishing feature of supersonic gas flows is the formation of discontinuity surfaces in face of obstacles. Across such surfaces the gas parameters change suddenly: velocity is suddenly decreased and density, pressure and temperature are suddenly increased. Discontinuity surfaces moving with respect to the gas medium are sometimes known as moving waves and steady discontinuity surfaces are called stationary shock waves. In this chapter we will consider the conditions of gas flow behind stationary shock waves. We will use the term "shock" for such surfaces. In the most general case the shock has a curvilinear form. Fig. 4.4.1, a shows the scheme of an attached curvilinear shock formed during interaction with a sharp body. Fig. 4.1.1, b shows the scheme of a detached curvilinear shock that appears before a blunt surface in a supersonic flow. An attached plane shock (Fig. 4.1.1, c) may appear during the interaction of a sharp body with plane faces.
b)
C)
normal
shock
Fig. 4.1 .l. Types of shock waves: aattached curvilinear shock; bdetached curvilinear shock; cattached linear shock.
From Fig. 4.1.1, b it can be seen that the surfaces of a shock may be oriented along the normal to the undisturbed velocity vector (angle of inclination 0,=z/2) or may be inclined at some angle 0, diverging from the right angle (Oc< 4 2 ) . In the first case it is known as a normal shock and in the second as an oblique shock. It is obvious that an attached curvilinear shock
137
can be regarded as a collection of a number of oblique shocks and a detached shodk may be regarded as one consisting of a normal shock and a system of oblique shocks. The formation of shock waves is based on the particular characteristic of the propagation of disturbances in a supersonic gas flow. By "disturbance" we understand the local irregularity associated with a rise in pressure. Such an increase in pressure appears in a flow during interaction with some obstacle known as the source of disturbance. Let us consider a source of infinitesimally small disturbances situated at a point 0 (Fig. 4.1.2). Such disturbances spread through the stationary gas (V=O) in all directions at the speed of sound a in the form of spherical waves in space and circular waves in a plane (Fig. 4.1.2, a). At time t the radius of wave r= a . t. If a subsonic gas flow ( V < a; Fig. 4.1.2, b) moves over the source the waves will be carried away downstream with the flow; during this the center of the waves moves with velocity V< a and the wave itself propagates at the speed of sound. After some time t the center of the waves will have moved by a distance V. t and the radius of the wave will be r = a . t, where a . t > Vt. Thus in a subsonic flow the disturbance spreads even upstream visavis the flow. In the special case of sonic speed (V=a) the leading front of the spherical or circular waves of the disturbance is restricted to the vertical surface or straight line tangent to the sphere or circle through point 0 respectively. This is because the distance by which the center of the waves moves in time t is equal to its radius at the same moment of time t. Assume that the oncoming flow has a supersonic speed (V>a). The center of the waves moves by the distance V . t in time t and the sonic wave propagates by distance a . t. Since a . t < V. t it is possible to draw a moving conical surface as an envelope of all the spherical sound waves (Fig. 4.1.2, c). This is called the cone of disturbance (Mach cone). On the plane of the moving circular waves we have lines of disturbance (Mach lines). On the Mach cone or Mach lines acting as the boundary separating the flow of the disturbed and undisturbed regions the disturbances are grouped more tightly because all sonic waves lie on this cone in the same phase of vibrations or phase of concentration. Such regions of disturbance, conical or plane waves determined by straight Mach lines are known as plane compression waves or Mach waves. The angle of inclination p for a conical wave or line of disturbance is determined from the condition (see Fig. 4.1.2, c) that sin p = a t/Vt; hence sin p = 1/M. (4.1.1)
The angle p is called an angle of disturbance or Mach angle. Supersonic flow carries all sonic disturbances downstream with the flow and hence limits their propagation by a cone or lines of disturbance inclined by the angle p.
1 39
The front of the plane wave propagates with the same sonic speed as the spherical (or circular) wave. Therefore the projection of the velocity vector of the undisturbed flow on the normal to the wave front is equal to the speed of sound (see Fig. 4.1.2, c).
Fig. 4.1.2. Nature of propagation of disturbances in a gas: agas at standstill; bsubsonic flow; csupersonic flow.
In a plane compression wave, as in a sonic wave, the gas parameters (pressure, density, etc.) change by an infinitesimally small value which, specifically, indicates the famous relation of physics for sonic speed a= ddpldp. In a disturbed region of flow the velocity remains practically the same as in an undisturbed flow. Therefore the plane compression wave is regarded as a shock wave of infinitesimally small intensity and in practice it can be assumed that the flow parameters across a plane compression wave remain unchanged. That is why this kind of shock wave is known as a weak shock and its leading front (Mach line) is known as a line of weak disturbances. It is natural to assume that the formation of a shock of finite intensity is connected with the superposition of plane compression waves which mutually strengthen the shock. Let us consider the process of generation of oblique shock wave. Assume that a supersonic flow initially moves over a smooth surface (Fig. 4.1.3). We will create a local increase in pressure at point A artificially by turning the flow through an infinitely small angle dB. This produces a plane compression wave C AB starting from point A as from the source of disturbance. It is inclined to the surface by an angle p. If an additional inclination by a small angle 6P is to be made, a new plane wave AC is formed = which begins from the same point A but is situated to the left of the Fig. 4.1.3. Formation of shock wave. first wave. However, in a supersonic flow, as shown earlier, the wave cannot propagate upstream. Therefore the wave AC will be carried away downstream until it coincides with the first wave. In this process a more powerful wave is formed which is con
140 AERODYNAMICS
siderably strengthened if the flow is turned some more. The shock wave of finite'intensity formed in this way has a propagation speed higher than the speed of sound at which the plane wave of disturbance propagates. Therefore this shock of finite strength must be inclined to the left of the plane wave AB and must occupy the position AD where it remains in equilibrium because its velocity of propagation will be equal to the velocity component of the undisturbed free flow velocity along the normal to wave front V sin O,, where 0, is the angle of inclination of shock. From the above discussion it follows that the angle of inclination of a shock of finite strength is more than the angle of inclination of the line of (cone) disturbance, i.e. 0, >,u.
141
sented in the form of a discontinuity surface for gas parameters, assuming that variations in these parameters take place abruptly. The problem of the shock lies in determination of the unknown gas parameters behind it from the given parameters characterizing the motion of the gas in front of it. For an oblique shock appearing in dissociating and ionizing gas there are nine unknown parameters: pressure p2, density p2, temperature T2, velocity V2, enthalpy h, entropy 5'2, speed of sound a2, average molecular weight pav2,and angle of inclination of shock 8, (or angle of flow inclination PC). Consequently it is necessary to construct a system out of nine equations. In these equations the parameters in front of the shock will be known parameters: pressure pl, density plyvelocity V1, etc. In place of velocity V2 behind the shock it is possible to determine its components Vn2and Vzz, normal and tangential to the shock respectively. As a result the number of equations needed goes up to ten. This system of equations will include the basic gas dynamic equations (equations of motion, continuity, energy and state), a series of kinematic relations for velocities and the thermodynamic relations characterizing the properties of the gas. Let us study each of the equations of this system. Figure 4.2.1 shows triangles of flowvelocities in front of the shock (parameters with index "1") and behind it (parameters with index "2"). The following relations for given components can be determined from this figure: VT2 V cos (8,  PC), Vn2 V2 sin (8,  PC). = 2 = The first equation of the system is determined from this: Vn2/VT2=tan (Oc 8,). (4.2.2) (4.2.1)
The second equation will be the equation of mass flow (continuity) which determines the amount of the fluid flow rate through unit surface area of shock per unit time:
Line of shock
I
Fig. 4.2.1. Schematic diagram of oblique shock.
142 AERODYNAMICS p l V n ~ = p ~ V;n ~ (4.2.3) where normal component of velocity in front of the shock (see Fig. 4.2.1) Vnl=V I sin 8,. The equation of momentum can be used for transformations of parameters across the shock. This will be the third equation of the system. It can be obtained by equating the variation in momentum of a fluid flowing in unit time through unit surface area of shock in the direction normal to it, to the pressure force: P ~ v i t  ~ V ; Z = P ~ . I P P (4.2.4) With the help of equation (4.2.3) this equation can be written in the following way:
Equation (4.2.4) could be expressed in the form P I + P I V ~ ~ = Pv~ 2 . P ~ z+ (4.2.4") The equation in this form represents the law of conservation of momentum during flow across the shock. If the change in the momentum of the fluid along a direction tangential to the surface is taken into account along with the zero value of the pressure gradient in this direction we get the relation
This fourth equation of the system indicates that tangential components of velocity across the shock do not change. From Fig. 4.2.1 it can be seen that V,I= V1 cos 8,; consequently equation (4.2.5) can be written in the form Vrz= V1 cos 8,= Vni/tan8,. (4.2.5') The equation of conservation of energy may be written in the form: i, (V:/2)= i2 (Vzl2). Under the condition that V ~ = V ~ + V V;= , Vi2+V:2 and V,,= V,,, ~ l this equation can be modified to (4.2.6) it t (V,21/2) i2+ (V22/2). = Combining the equation of state before and after the shock we get the relation
The four equations useful in determination of enthalpy, entropy, average molecular weight and speed of sound in the dissociating gas of the system we are considering can be represented in the form of general relations of these parameters as functions of pressure and temperature: i2=f1 ( p2, T2); (4.2.8) S =f2 ( ~ 2T2); 2 , (4.2.9) ~ a v 2 = f 3 (PZ,T2); (4.2.10) (4.2.1 1) a2 =f4 (~ 2 T2). , These functions are analytically not expressed in the usual form but they are determined by means of experimental research or with the help of fairly complicated calculations from the solution of the corresponding thermodynamic equations. The relations for the above functions are usually prepared in graphic form and their values are tabulated in special tables of the thermodynamic functions of air at high temperatures (see [7], [17], [18]). Let us express the basic parameters behind the shock in terms of relative change in normal components of velocity, i.e. by the quantity AVn=AVn/Vnl =(Vn1 Vn2)/Vn1. From (4.2.3) the density ratio will be p2/p1=1/(1 dr?;t), and from (4.2.4') the pressure ratio will be (4.2.13)
(4.2.12)
Introducing the concept of a "normal component" of Mach number MI in the above condition so that Mnl=Vnl/al or Mn1=M1 sin 9, and assuming that the gas remains undissociated before the shock with sound velocity a1 = 6k1~11p1, get we (4.2.15) ~ 2 1 = 11 k,Mtn AV, ~ + in place of (4.2.14). The enthalpy ratio izlir can be determined from (4.2.6): i2/i1= 1 + (112 il) (Vzl  Vi2). The difference between the squares of the velocities will be
To determine the temperature ratio T2/T1we use the equations of state for the flows in front of and behind the shock. This leads to
Replacing the ratios of densities and pressures by equations (4.2.13) and (4.2.15) respectively, we get T2/Tl=(I +kl M,21AVn> (1 A pn) (Pav2 IPavl). To determine the velocity behind the shock the relations (4.2.17')
v$=v:+
are used, from which
Vi2 and
v:=
V:+ v;,,
Let us determine the relation for the angle of inclination of flow behind the shock. From (4.2.2) and (4.2.5') we have AT, = 1  [tan (0,  Pc)]/tan 0,. Hence, taking into account that tan (6,  6,) =(tan 6, tan Pc)/(l +tan 8, tan PC), the relation for PCwill be (4.2.19)
The change in velocity dl/,,is determined from (4.2.13) in terms of nondimensional density: dVn= I(PI/P~). (4.2.21) From this it follows that the pressure ratio, temperature ratio, enthalpy ratio and angle of inclination PC can be expressed as functions of density ratio p2/p1. Besides, the quantities Vnl and Mnl can be replaced by the values VnI=Vl sin 8, and M,I=MI sin 8,. So the solution of the problem of an oblique shock with a known angle of inclination 8, leads to the determination of density ratio pz/pl or function
145
AV,. This function can be determined with the help of relations (4.2.16) and (4.2.17'), each orwhich can be rewritten in the form of a quadratic equation of A%. From the first quadratic equation we find
AV, =A s ( A ~ B ,
where
The signs before the quantities under the square root, minus in (4.2.22) and plus in (4.2.23), reflect the condition that velocity is always less behind the shock than in front of it and hence AT,< 1 must be observed. Here the plus sign in (4.2.23) also accounts for the physically possible value of AV, c 1. Equations (4.2.22) and (4.2.23) are solved by the method of successive iterations. The problem of an oblique shock can be solved if angle PC is given. Then, the calculation of angle 0, is carried out with the help of relation (4.2.20), from which tan2 0,tant?, AV, 1 $   w =o. tan PC' 1 AV, 1 AV,
One solution (plus sign before the square root) gives a higher value of angle O,, which occurs in a detached curvilinear shock, and the other (minus sign before the square root) gives a lower value of O which is observed in an , attached shock of fairly weak strength. For ease of calculation it is desirable to calculate 0, in advance on the basis of given values of 8, and to prepare the corresponding table or graph. With the help of this table or graph for some value of AE corresponding to the values of ratios pz/pl, p2/p,, etc. obtained earlier it is possible to find angle 0, or PCwith one of them as a known quantity. The calculation of velocity V l and Mach number MI of the flow upstream of the shock, where the parameters in the shock are observed for given values of angle 8, (or PC) and V,1 (or M,I), is performed with the help of the formulas
146 AERODYNAMICS
2.2 Normal shock Formulas for calculation of flow parameters across a normal shock can be obtained from the above relations for an oblique shock by taking Oc = n / 2 and 9 0 , Then we get velocity V,I =V1 and Mach number M,I = M I . In swit,= . ching from oblique shock to normal shock it is required to exclude index "n" in the above relations for oblique shock. Then the basic relations take the following form:
pz/pl = 1 +klM?AV; i2/il= 1 (V!/2) (A57/il)(2 AT); TJTl = ( I +kt M:A@ (1  A v ) (Pav2IPavl); V;/V:=l AV(2AT), where the change in relative velocity is determined with the help of relations (4.2.22)(4.2.24): dV=lJ1[2 (i,i,)/V:]; (4.2.32) (4.2.33)
AF=A+JA~B;
AT
(4.2.35) A 1 ( P I / P ~ ) These are the general relations for shock waves. Let us now analyze the nature of motion with the help of these relations and study the methods of calculation of gas parameters behind shocks in the case of constant specific heats. Then we will discuss in detail the practical methods of calculation of similar parameters for a dissociating medium, i.e. in the more general case of variable spec& heats.
3. Oblique Shock in a Gas Flow with Constant Specific Heats
v=
One problem concerning a gas flow behind a shock wave with constant specific heats (c,,, c,) happens to be of great theoretical and practical interest. Although such a flow is seen as a particular (ideal) case of the motion of gas, whose physicochemical properties change to a greater or lesser extent during transition through a shock, the solution of this problem allows us to establish
147
a qualitative picture of flow through a shock. The results to be obtained in the form of the usual relations characterizing the variations in gas parameters during motion through a shock can be used for an approximate estimation of these parameters when the more general case of heat flow parameters is considered. Finally, the problem we are examining has its own significance because its solution is directly applicable in the determination of the gas parameters behind a shock occurring in a flow of comparatively low supersonic speed at which variations in specific heats in a compressed gas are negligibly small. These velocities of flow, to be determined for the most intensivenormalshock, correspond approximately to Mach number M, < 5 to 6.
3.1 System of equations The method of calculation of an oblique shock to be considered here is based on application of a system of equations which happens to be the particular case of the system of equations (4.2.2) through (4.2.1 1). If the specific heats remain constant during the flow of gas through the shock it follows that average molecular weight remains unchanged and the speed of sound and enthalpy depend only on temperature. Accordingly the equations (4.2.8), (4.2.10) and (4.2.1 1) take the following form:
i2 = cpT2;
Llav2 = Pavt = ,&v
=const; a:=kRT,.
In place of equation (4.2.9) it is necessary to use the thermodynamic equation for entropy of an undissociating perfect gas: dS =(di/T)  (dp/pTj. Recalling the equations di = c, dT and dp = d (RpT) = RpdT+ RTdp , we get dS=c,d In TRdln TRd In p. But cpR = C, and Rlc, =(c,  c,)/c, Consequently dS=c,[dIn T(kI) dln p]. Integrating this equation with k const, we find S= c, In (T/pkl) +const. Replacement of T by its value from the equation of state p = RpT leads to T/pkI =p/pkR and In ( p/pkR)=ln (p/pk)1n R.
=k  1.
Absorbing In R in the constant on the right side of equation (4.3.4), we get the relation for entropy in the following way:
Transferring this equation to the flow conditions before and after the shock and determining the difference in entropy on either side of the shock, equation (4.2.9) can be replaced by the equation used in the theory of oblique shock: (4.3.6) In [(PZ/PI)( P:/P%I. The entropy behind the shock within the accuracy of the constant can be determined by an expression
s 2  s = c, 1
The equation of state (4.2.7) in the case of constant specific heats can be simplified to
In all these expressions the variation in relative velocity ATn is the known quantity. Let us determine this quantity assuming that the angle of inclination of shock 8, and the Mach number of oncoming flow MI are known quantities. To do this we use equation (4.2.4'). Dividing this by plVn1= p2Vrz2,we get
Introducing here substitution of p2/p2and p,/p1 by a:/k and az/k respectively and using equation (3.6.21) for the speed of sound, we find
: : : Substituting here the values of V,2= V,22+ V and V= V$ + V and multiplying both sides by VnlVn2, we obtain after some simplifications the following expression:
Excluding here the trivial solution of Vnl Vn2=0 that corresponds to the case of absence of shock, we get an equation
which helps in the determination of the velocity of flow behind the shock. Equation (4.3.10) is known as the basic equation for an oblique shock. From this equation the relative variation in velocity ATn=(Vnl Vn2)/Vnlcan be in found. For this purpose we express VnlVn~ the form
where A1 = Vl/a*. Replacing A1 by the formula (3.6.23) and performing further transformations, we get
Let us introduce the notation 6 =(k  l)/(k + 1). Then formula (4.3.11) will be
Substituting this quantity in (4.2.13), the density ratio can be determined as p2/pl=Mf sin2Bc/(l 6 +6 M: sin20,). (4.3.13) Let us take Formula (4.2.15) can be used to find the pressure ratio pzlp~. Mn,=Ml sin 9, here and in place of AV, the relation (4.3.11'). Further, we replace quantity k by (4.3.14) k = (1 +6)/(1 6) according to (4.3.12).
150 AERODYNAMICS
As a result we get
The intensity of shock is characterized by this quantity. For this purpose it is also possible to use the relation One more parameter characterizing the strength of the shock and, in particular, the coefficient of pressure = (p2pl)/ql, where q, =[(I +6)/(1 6)] p, ~ : / 2 can be obtained with the help of formula (4.3.15'). Representation , of (4.3.15') in the form o f j 2 with respect to quantity ql gives
Eliminating the quantity M: sin2 8 from equations (4.3.15) and (4.3.13), we , get the relation between expressions for pressure and density, i.e. a Hugoniot equation:
This relation is also known as the adiabatic shock relation which, unlike the usual adiabatic (isentropic) equation of the type p= Apk, determines the variation in flow parameters across the shock wave. This process of transition through the shock is accompanied by a rise in entropy which can be determined by equation (4.3.6). In this way the process of gas flow through the shock is nonisentropic. With the change in the above parameters the temperature behind the shock will increase. Its value is calculated by the equation of state:
The peculiarity of flow across the shock, unlike isentropic flow, consists in the different character of the variations in the gas parameters. From (4.3.15) and (4.3.16) it follows that as MIco the pressure and temperature rise indefinitely. At the same time equation (4.3.13) shows that under the same condition of M 1  m the density tends to some finite value which is equal to ( ~ 2 / p l ) = 116. At k= 1.4 this gives the value of 116=6 . From formula ~~+~ p = Apk or T= Bp kl ( A and B are constants) it follows that an infinite rise i n density and temperature corresponds to an infinite increase in pressure in the case of an isentropic (adiabatic) process. From this it may be concluded that for the same variation in pressure in a flow across a shock and an adiabatic process the first is characterized by more intensive heating of the gas, which leads to some reduction in density.
Formula (4.3.16) determines the ratio of the squares of sound velocities in front qf and liehind the shock in the following way: Using this relation and an expression (4.2.18') it is possible to determine Mach number M2 behind the shock. Substituting the expression (4.2.21) for A K in (4.2.1 8'), we get Taking (4.3.18) and (4.3.17), the ratio of the left sides and right sides gives the ratios of the squares of the Mach numbers
where Tl/T2 and pl/pz are found from the formulas (4.3.16) and (4.3.13) respectively. The formula for calculating M2 could be obtained in somewhat different form, with the help of the equation of momentum (4.2.4"). Let us write this equation in the form:
Recalling that kp/p = [(I+ 6)/(1d)] p/p and determining p2/p1 by the Hugoniot equation (4.3.13'), we get
: The quantity M .sin2 0, can be replaced by equation (4.3.13), so that Let us determine the stagnation pressure for flow conditions behind the shock. Taking the gas flow behind and in front of the shock as isentropic, the thermodynamic relations wherep, andp;, po and pi are stagnation pressure and density in the regions of flow in front of and behind the shock respectively. From these relations the coefficient of established pressure in the shock will be v ~ = P ~ / P ~ = ( P z / P ~ )~(P~IPO)~. (~11 2)~ Multiplying both sides of this equation by (p~/p,)~, get we
We recall the equation of energy (V2/2)+cp.T=const and write it down for t'he conditions in front of and behind the shock: (V:/2) + cpT1 =(V:/2) cpT2. At points of complete stagnation Vl = V2=0. From the equation of energy it follows that the temperatures at the two points are equal, i.e. T,=T;, or in other words po/po=p~/p~. Hence,
where p1/p2 and pz/pl are obtained from formulas (4.3.15) and (4.3.13) respectively. According to (3.6.28) the stagnation pressure will be
hence P1
PO = [(l +6) M:
Then the coefficient of stagnation pressure behind an oblique shock will be:
f
Analysis of relation (4.3.20) shows that the pressure ratio pd/po behind a shock of finite strength is always less than unity. A more intensive shock results in higher loss of stagnation pressure and hence the ratio p&p0 diminishes more sharply. Let us take a stream line and two points on it such that one is situated before the shock and the other behind it (in the general case at some distance from the wave front). Let the velocities at these points be the same and
equal to V. As can be seen from (3.6.26), the pressure at the points before and after the shock will be and
v 2
(1 Km !x
kl(k1)
respectively. From expression (3.6.15) it follows that, since po/p,=p~/p~, the values of maximum velocity are equal, i.e. Vmaxt = Vmax2= Vmax. Thus in the two formulas for p(1) and p(*)the quantities in brackets are equal. Therefore pressure pcl) before the shock will be more than pressure pC2) behind the shock because p;<po. This is how the static pressure loss comes about. In explaining the physical nature of these losses the shock should not be regarded as a discontinuity surface. It is necessary to assume that the real process of compression takes place in a layer of small thickness of the order of the mean free path length of the gas molecules. Only this kind of flow process through a shock is possible: physically it is incorrect to posit the juxtaposition of two regions exhibiting finite differences in temperature, pressure and density. This is no more than a mathematical abstraction. The process of flow across a shock of small thickness is characterized by such large gradients of velocity and temperature that there will be considerable friction and heat transfer effects in the compression zone. From this it foI1ows that the irreversibIe losses of kinetic energy by the gas during flow across the shock are connected with frictional forces and also with heat conduction. The effect of these dissipating forces and the heat flow inside the compression zone is an increase in entropy and consequent decrease in static pressure in the flow behind the shock in comparison with the isentropic process of compression.
3.3 Angle of inclination of oblique shock Parameters behind an oblique shock are determined not only by Mach number MI but also by angle of inclination 0, of the shock. Its value, which depends on Mach number MI and angle of inclination PC of the flow, can be determined with the help of equation (4.2.19). In this equation the replacement of quantity A% by its value from (4.2.21) gives relation
tan Oc/tan (Oc PC) pzlpl . = Determining pl/p2 from (4.3.13), we find MI sin20, tan 8, tan(@,PC) 16+6M:sin2OC On the other hand this relation helps us to find angle of inclination PCof the flow behind the shock from the known inclination of shock. The connection between angles Oc and PCfor various values of Mach number MI is shown graphically on Fig. 4.3.1. As can be seen from the graph the value of angle PC
154
AERODYNAMICS
changes with the decrease in angle 0,. In the region to the left of the dotted line indicating the maximum value of angle PC (PC) the value of angle Oc diminishes. Conversely, on the right side of this dotted line Bc increases. This type of relation is based on a different form of shock. In the first case the variation in angle 0, corresponds to the attached shock of curvilinear form in front of a sharp body. With greater bluntness (angle of leading edge) the angle of flow inclination increases and hence the angle of inclination of the shock increases. The maximum angle of flow inclination PC= PC is determined only by the given Mach number MI. This angle is also known as the critical angle PC, of inclination (or turning) of flow. The points corresponding to the values of this angle are connected by the dotted line in Fig. 4.3.1. Experiments show that the flow behind an attached shock is steady in the sense that its form is maintained as long as the angle of flow inclination in the whole region remains less than critical. Accordingly this kind of flow is called a subcritical flow.
Fig. 4.3.1. Variation of angle of flow inclination with angle of shock Bc for different Mach numbers MI.
With any further increase in the angle of the leading edge the angle PC could be critical. With a still higher angle of leading edge steady flow behind an attached flow cannot be realized. The shock moves ahead from the nose of a body. A steady low region, characterized by inclination through an angle less than the critical value, comes into being behind this detached shock. However, as distinct from a subcritical flow, this flow is called supercritical. This definition corresponds to the fact that the angle of the nose part of a body in a flow surpasses the vaiue\at which an attached flow may still be observed. A detached shock completely changes form, which is clearly seen in the example of a flow past a sharp cone or wedge (Fig. 4.3.2). As long as the flow is subcritical the shock remains attached to the body and the surface so forrn
155
ed is rectilinear. A flow around thick wedges or cones may be supercritical and then the shock separates from the body and adopts a curvilinear form. At the point of cross section of surface of the shock to the axis of flow the angle of shock inclination is 0c=;n/2 and hence the flow parameters change their value by the rule of normal shock. There exists a region near the axis where some part of the shock is practically normal.
As we move away from the axis the angle of inclination 0, decreases, as shown in the graph in Fig. 4.3.2. But it maintains higher values in some places, corresponding to subcritical flow. The variation in the angle of flow inclination has the opposite character. At the peak of the detached wave behind the straight part the value of angle PC=0 and then it increases. At some point on the wave surface the angle PCbecomes critical and then it decreases again along with the angle of inclination of the wave. Here, as can be seen from (4.3.25), the angle Oc tends to a value Bc= y=sinJ (l/Mr) with the limit of Pc+O. So the angles of inclination of shock for given values of Mach number MI vary in the range of p,<Oc<n/2. The value of Oc=,u corresponds to a weak shock representing a simple disturbance wave. For a curvilinear shock (see Fig. 4.3.1) it is possible to find two points corresponding to the two values of Oc which determine the single value of PC at a given MI. This angle is calculated by formula (4.2.20) which, after substitution for the value of AT*from (4.3.1 I), takes the form tan
Here the strong shock corresponds to the higher value of angle 8, and the weax shock with supersonic flow behind it (if we exclude the region around the shock with angle PCnear its critical value, where the flow can be subsonic) corresponds to the lower value of angle Oc. Calculation of the angle of inclination of an oblique shock can be carried out with the help of formula (4.2.25). Replacing AE by the density ratio according to (4.2.21), we get
On each of the curves in Fig. 4.3.1 there is a point where the value of Mach number M2= 1 behind the curvilinear shock. Joining these points with a continuous curve, we get the boundary of two regimes of flow behind the shock: on the left side of this curve the flow is supersonic (M2 > 1) and on the right side subsonic (M2< 1).
4. Hodograph of Velocity
An analytic solution of the problem of determining flow parameters behind an oblique shock was set forth in the preceding section. There exists a graphic method along with this analysis which is based on the principle of the hodograph of velocity. A hodograph of velocity is a curve representing the geometrical position of the ends of velocity vectors behind a shock. Let us study the equation of a hodograph of velocity. Let point A (Fig. 4.4.1) be the end of velocity vector V2 and be located accordingly on a hodograph velocity that is constructed in the system of coordinates with its horizontal axis coinciding with the direction of velocity Fj7, front of the shock. Hence an inclination of in velocity vector 'F2 determined by angle 8,. Let the horizontal and vertical is components of velocity behind the shock be denoted by uz and wz respectively. From Fig. 4.4.1 it can be seen that 242 and w2 can be expressed by the components V 2 and V of n , velocity V normal and tangential to 2 the plane of shock in the following w, way:
u2 = V ,
.
02
w2 = V,
'
"'
The component Vn2is determined (4.3.10), where V n l = V 1sin 0,,V,= VI cos Bc. Accordingly
To eliminate angle Bc we use equations (4.4.1). Multiplying the first equation , of (4.4.1) by cos 8, and the second by sin 8 and summing them, we have
= V, .
This equation, connecting the variables w and u2, represents the equation of 2 a hodograph of velocity. Usually it is written in the form
Let us introduce nondimensional parameters ,Iu= uz/a*, I , = wzla*, 1 1 =Vila" and 6= (k l)/(k+ 1). Then the hodograph equation can be written in the following way:
Equation (4.4.4') is reproduced on plane A, ,Iu graphically by the curve , known as the strofoid (Fig. 4.4.2). Let us determine some characteristic points of the strofoid. Specifically, let us find the coordinates of points of intersection A and D of the strofoid and axis lu.From (4.4.4') it can be seen that the condition of A,,,= 0 is fulfilled if AU=1 or A, = 1/11. The value ilu =1 1 1 determines the coordinates of point A and gives the solution for an infinitesimally weak shock across which the velocity does not change. The other value AU= 1/11 determines the coordinate of point D nearest to the origin and happens to be the solution for normal shock (see Section 5). From the construction of the strofoid it follows that its two branches on the right side of point A extend to infinity asymptotically to the line passing through point B and parallel to the vertical axis. The coordinate of this point
can be obtained from (4.4.47, recalling the limit of A,,+co. As a result we get the Condition (1  6) A:+ 1 1,1, =0 from which the coordinate of point B can be obtained:
n,=nl
(1  s ) + ( i l n , ) .
Any point on the branch of the strofoid extending to infinity in principle gives the solution for compression shock. For example, let us examine point F i n Fig. 4.4.2. It may be assumed that for a shock behind which the velocity direction changes by the given amount of angle PC the velocity increases due to shock to the value 12 determined by the length of line OF. Meantime the density and pressure can be reduced due to shock. In other words, in a given case there can exist expansion shock in place of compression shock. However, formation of such a shock is physically impossible. To prove this we use formula (4.3.6) for variation in entropy. Applying the relations p21p';= P ~ / ( ~ l~ P ~~= ,p o / and recalling that pi/po= pi/Po, from formula (4.3.6) p / ) P~ we get
S,  S, =R In
=  R In v,.
(4.4.5)
In the case of a compression shock pa > p i and hence, S, S , > 0. This conclusion corresponds to the second law of thermodynamics, according to which the entropy of an isolated system of gas having jumps must increase. Let us now consider a reverse flow where a gas in condition (2), characterized by stagnation pressure p i , moves to the condition (1) with stagnation pressure po by means of an expansion shock. 1n this case, by analogy with (4.4.5), the change in entropy
is against the second law of thermodynamics. From this it follows that an expansion shock cannot appear. From the above discussion it can be seen that the process of motion through a shock, which happens to be adiabatic by nature, as it occurs in a thermally isolated system, will represent by itself an irreversible adiabatic nonisentropic process. It can easily be checked with the help of (4.3.6) that the real process of rise in entropy (5'2S1 >0) corresponds to the case of a supersonic flow [MI > 1 (normal shock), M1 sin 8, > 1 (oblique shock)] and the physically impossible phenomenon of reduction in entropy (S2S1< 0) corresponds to the subsonic flow (MI < 1 and M I sin 8, < 1). Thus compression shocks can take place only in a supersonic flow. It should be emphasized that the relations obtained for variation in entropy IS2SlI are valid for the case of an irreversible process of motion through a shock accompanied by an isentropic gas flow up to the shock and also behind it. From the above discussion it follows that the branches of the strofoid proceeding to infinity have no physical meaning. The remaining part of the strofoid (left side of point A) has physical significance: it is called the shock polar. This kind of curve is drawn for given It (or Mach number MI). Some such curves drawn for different values of I represent a family of shock polars 1 which help to find the velocity of flow behind a shock and its angle of inclination graphically. Take an attached shock at a wedgeshaped surface with semiangle /Ic (see Fig. 4.1.1, c). To determine the velocity behind this shock for a given value of 1 1 (or MI) a shock polar can be constructed and a line can be drawn through point 0 at an angle PC(see Fig. 4.4.2). The point of intersection N of this line with the shock polar determines vector ON, whose modulus gives the value of relative velocity 1 2 behind the shock. According to formula (4.4.3), which can be expressed by tan 0, = (I1  Au)/1.,,,, (4.4.3') the angle ANG on a shock polar is equal to the angle of inclination 8 of , the shock wave. It is easy to see that this angle can also be determined as an angle between the horizontal axis and the normal to a line joining the ends of the velocity vectors before and after the shock (in Fig. 4.4.2 the corresponding points are A and N). From the shock polar we arrive at the conclusion that with a reduction in angle PC (point N moves along the curve toward point A) the angle of inclination 8, of the shock decreases. In the limit of /Ic+O the point N approaches point A, which physically corresponds to the transformation of a shock wave into a shock of very low strength, i.e. into a line of weak disturbance. The angle of inclination of this shock Bc=p is determined as an angle between the horizontal axis and the line perpendicular to t h ~ tangent of a shock
polar at a point A (see Fig. 4.4.2). 'An increase in the angle of inclination of the flow (in Fig. 4.4.2 this corresponds to a shift of point N away from A) leads to an increase in the angle of shock and a rise in its intensity. On the shock polar it can be seen that at some value of angle 8, the line drawn through point 0 touches the curve at a point C . The angle of inclination of this tangent line determines the maximum angle of flow inclination, which we earlier designated the critical angle (PC PC,). Let the angle PC &. On the curve this angle corresponds to the = > dotted line drawn from point A and does not intersect the shock polar. Thus >PCo is not possible to find a solution for shock graphically with the it at PC help of the shock polar. This is because the condition PC> PC,does not correspond to the assumptions of the rectilinear type of shock attached to the nose of a surface, on the basis of which the shock equations were derived. Physically speaking, the case of an angle of wedge PCproducing a more than critical angle of turning results in a detached curvilinear shock. The determination of the shape of this curvilinear shock in front of a body is material for a special problem of aerodynamicsrelated, in particular, to the conditions of the supersonic flowinteraction of wedgetype bodies. If this problem is not solved it is possible at least to give a qualitative estimation of the variation in parameters in some region ahead of the body with the help of the shock polar from point D to A. When the form of a shock is determined for given conditions of flowinteraction (apart from calculations this can be done with the help of the aerodynamic wind tunnel) it is possible to establish the quantitative relation between the points of the shock polar and the shock surface. For example, let angle PCbe given and the points E and N lie on the shock polar. The angle of shock BcN=LAhJG corresponds to point N and angle OCE= L A E K (EK L OB) corresponds to point E. If the configuration of the wave front is known it is possible to find on it point N' with an angle of inclination of wave OcRr and point E' with angle BcE (see Fig. 4.3.2) by direct measurement. In the same way point C' corresponding to the critical (maximum) angle of turning PC,can be found. Point D on the shock polar corresponds to the peak on the given surface of the detached shock (normal shock) and the end point A of the polar corresponds to the farthest part of shock, which transforms into a line of weak disturbance. For an attached shock (PC<Per) two solutions can be obtained, as seen from the shock polar. One of them is point E, which corresponds to a lower velocity behind the shock. The other, point N, refers to a higher velocity behind the shock. It is found experimentally that attached shocks with higher speeds behind them are realized physically, i.e. shocks of low intensity occur. If the arc of a circle of radius equal to unity (in dimensional axes wz, 2.42 this refers to a radius equal to the critical speed of sound a*) is drawn on the
graph regions of flow can be determinedsubsonic and supersonicthat are related to the points lying on the shock polar to the left and right of the arc. The part of the flow representing subsonic speed is shown shaded in Fig. 4.3.2. From the shock polar it can be seen that the flow velocity behind a normal shock is always subsonic. At the same time the flow velocity behind an oblique (curvilinear) shock can be supersonic (corresponding points on the shock polar lie on the right side of point S) as well as subsonic (the points on the shock polar lie on the left side of point S). Further, the points lying between S and C correspond to an attached shock behind which the velocities are subsonic. Experimental research shows that the shock remains attached for wedge angles PC less than the critical angle PC, or more than LSOB. However, in the latter case it becomes curved. Here the theoretical values for angle Qcand gas velocity A2 in the whole region behind this curvilinear shock, obtained from the shock polar on the basis of angle PC do not correspond to the real values.
5. Normal Shock in Gas Flow with Constant Specific Heats
The appropriate relations for a normal shock are obtained from the condition that 0,= 7r/2 and hence V , 1 = V1 and Vn2= VZ. The basic equation (4.3.10) takes the form v~Vz=a*~ (4.5.1) or A1Az= 1, (4.5.1') where 11= Vila*, Az= Vz/a*. The relative change in velocity is found from (4.3.1 1'):
For the ratios of densities, pressures and temperatures, the corresponding relations can be found from (4.3.13), (4.3.15), (4.3.16): (4.5.3) pzlpl =M:/(l 6 6 Ma;
Eliminating M: from (4.5.3) and (4.5.41, we get an equation of adiabatic flow for normal shock which does not differ in external form the equation for oblique shock [see (4.3.13')l. Taking Bc=.n/2 in (4.3.19), the relation for the Mach number behind a normal shock can be obtained:
Let us examine the gas parameters at a stagnation point (at a critical point) of a blunt surface situated behind a normal shock (Fig. 4.5.1). Pressure pi at this point is determined by the formula (4.3.20') in which the ratios pzlpl andpzlpl are found from (4.5.3) and (4.5.4) respectively. Incorporating this, we have
(4.5.8)
Knowing absolute pressure p; the nondimensional quantity jo= (p; pl)lql, i.e. the coefficient of pressure at the point of complete stagnation, can be determined. Recalling that the velocity head
we have
In Section 3 it was proved that the stagnation temperature behind the shock does not change, i.e. T; =To. Consequently at a point of complete stagnation
v,
(4.5.10)
\
Fig. 4.5.1. Gas parameters at point of complete stagnation behind normal shock.
The expression ii = cpT; and il = cpT, can be used to determine enthalpy. Accordingly, at the point of complete stagnation i;=il
(I+M:). 16
(4.5.11)
6. Shock at Very High Supersonic Speeds and Constant Specific Heats of Gas
At very high (supersonic) speeds with values of M I sin 0,% 1 the nondimensional gas parameters behind the shock are close to their limiting values
obtained at M I sin 0,+co. From (4.3.11') it follows that under this condition
Inserting this value in (4.3.27) in the limit of M I sin 0,+co, we get tan 0, = (cot Pc/26) [l 6 5 d(1 612 4 6 tan2 PC]. (4.6.3)
Let us find the limiting value of the coefficient of pressure. For the conditions immediately behind the shock it follows from (4.3.15") that at M I sin 0,+m and M1+co
p 2 =2
(1  6) sin2 0,
(4.6.4)
For the stagnation point the corresponding value is obtained from (4.3.23):
The ratio of coefficients of pressure will be In a particular case when 6 = 116 (k = 1.4) the ratio >,/j2 = 1.09. The limiting value of Mach number M2 may be found from (4.3.19), taking into account the relation (4.3.16) for T ~ / T I . Going back to the limit, at M I sin B,+coand MIco we get Mg=[1/6 (1 +a)] (cot2 0,+d2). (4.6.7) To find the limiting parameters behind a normal shock by the above relations it is necessary to take 0,=n/2. Then from (4.6.4) and (4.6.5) we get
will be the same as for an oblique shock. The It is clear that the ratio Po/& limiting Mach number behind a normal shock
17 For 6 = 116 (k = 1.4) the Mach number M2= 4 / x 0.38. The actual values of nondimensional parameters behind a shock with finite but very high Mach numbers will depend on MI. Let us study the relations for the case where attached shocks appear before thin wedges and are therefore inclined at low angles. Assuming that and tan 0 , ~ ~ 0 , tan (0,PC) wB,p, in (4.3.25), we get
164
AERODYNAMICS
ec ec
Solving this equation for 0,//3, and assuming that the physically possible condition will be 8,/Pc > 1, we find
Examining equation (4.6.9), it can be seen that the parameters determining flow for a shock at very high speeds are combined in functional groups that describe the solution of the problem of shock in a wide range of Mach numbers M, and of values of PC. Equation (4.6.9) is an example of a similarity relation. From this equation for the ratio ec/PC is found that the quantity it K=M& happens to be the similarity parameter. This similarity is to be understood in the sense that the ratios of angles 0c//3c will be the same for flows with identical parameters K independent of the absolute value of the quantities that characterize hypersonic flows. As K+co the ratio Bc/Pc tends to the limit equal to
Let us now take the relation for the coefficient of pressure. At small values of Oc the formula (4.3.15") takes the form
(4.6.11')
This formula shows that K also happens to be the similarity parameter for the ratio j2//3f. As K+co the value of this ratio will be
165
According to (4.3.13) the density ratio at small values of 8, could be ' expressed by pzlp t = KZJ(1 6 6 K3,
(4.6.14)
From (4.3.18) it follows that the second term on the right side can be neglected at small 1, (sin OCwOc);hence V2w Vt. Bearing this in mind, the ratio of the 9 squares of Mach numbers using (4.3.19) will be
M;/M?= Tl/T2. Substituting the ratio TlIT2 here by formula (4.3.16) and taking sin 0cx8,, we get
7. Solution of Problem of Shock in a Gas Flow with Variable Specific Heats Taking into Account Dissociation and Ionization The parameters of air at some altitude H (pressure pl, temperature TI, density pt, etc.) and the value of normal velocity component V,I are selected as the initial data in solving the problem of a shock wave in dissociated and ionized gas. In this instance an oblique shock, in a given case, is regarded as
7;/h
40
30
2 0
10
Fig. 4.7.1. Ratio of air temperatures behind and ahead of shock taking into account dissociation and ionization:
thick lineTI = 220K; dotted lineTI =350K.
a normal shock. Assuming the value of d i b 0 . 9 to 0.95 as a first approximation, which refers to the density ratio across the shock
we find pressure pz from formula (4.2.15) and enthalpy i2 from expression (4.2.16) which is, obviously, near the stagnation enthalpy ii. Then, from the i S diagram [7, 171 temperature Tz and from Fig. 1.4.7 average mole6 /4 20 cular weight pavz can be determined. In place of these diagrams 15 appropriate tables of thermodynamic functions of air [la] can be used. This increases the accuracy 10 of calculations. The density p2 can be found 0 from the values of p2, Tz, pav2 obtained and the equation of state (1.4.8). The value of A can be Fig. 4.7.2. Ratio of air densities behind and ahead of shock taking into account dissociaimproved by formula (4.2.21). tion and ionization: Then with this value the pressure tlzick line TI =220K; dotted liiteand enthalpy can be obtained by T~=350K. formulas (4.2.15) and (4.2.16) respectively as a second approximation. With the help of these more accurate values of temperature and average molecular weight can be obtained from tables and graphs. Knowing the more accurate values of p2, T2 and p,2, the density in the second approximation can be found by the equation of state. This iteration process is continued until the required accuracy is achieved. An oblique shock can also be studied on the basis of known parameters of undisturbed flow (including Mach number MI) and the angle PC. As a first approxi P 2 / 4 800 mation the angle of shock 0, for an undissociated gas is obtained [(see (4.3.25)] 600 and then the necessary corresponding p2 quantities AE, and i2 can be calculated rroo by formulas (4.2.19), (4.2.15) and (4.2.16). Using these values the temperature T2 and 200 average molecular weight paV2 calculatare 0 ed from tables [18] and graphs [7, 171. 5 10 15 20 Mnr Further, with the help of formulas (4.2.23) Fig. 4.7.3. Ratio of air pressures and (4.2.24) the value of A Y , can be imbehind and ahead of shock proved and from the expression (4.2.19) taking into account dissociation tan Oc and angle 0, are rendered more and ionization.
167
accurate. The remaining parameters are corrected by the respective formulas. The calculations of gas parameters behind normal shock are similarly performed with the use of tables and graphs of thermodynamic functions at high temperatures. For this we must consider Oc=a/2 and pC=Oand then relations (4.2.27) through (4.2.35) should be applied. In the presence of dissociation and ionization the relative values of gas parameters behind a shock wave depend not only on temperature, which is the characteristic of variable heat coefficients, but also on pressure. These relations are graphically shown in Figs. 4.7.1 through 4.7.3. The calculations of temperature ratio and pressure ratio were carried out for the average values of temperature TI 220 and 350K, which are the pro= bable minimum and maximum values chosen according to the change in temperature with altitude in the case of low and high yearly mean values. The data obtained show that dissociation and ionization indicate significant variations in equilibrium temperature and density in comparison with the case of constant specific heats (k= 1.4=const). So far as pressure is concerned it depends very slightly on physicochemical changes in the air. The ratio p ~ / p ldiffers negligibly from its maximum value pz/pl = 1+kl M ~ I which is determined only by the ~onditions undisturbed flow and not by of the change in the structure and physicochemical properties of air behind a shock wave. The calculations also show that at one and the same angle in a real heated gas medium the inclination of flow (angle PC)is more than in a perfect gas (k=const). This leads up to the fact that in a heated gas a detached shock appears later than H ,km in a cold gas. Specifically the wedge angle at which detachment of shock begins is greater in a heated gas than in a cold gas. Calculations of parameters at the point of complete stagnation: The theory of normal shock has important practical applications in the determination of gas parameters at the point of complete stagnation. This is realized in the following way. Entropy S z is found from the values of i2, p2 obtained with allowance for dissocia2000 4000 6000 v,, m/sec tion and ionization from the iS diagram or the thermO Fig. 4.7.4. Stagnation pressure and temperadynamic functions of air. Assumture at point of complete stagnation.
ing that the flow behind the shock is isentropic entropy Si at the stagnation point can be taken as equal to the value S2 behind the shock wave. Also at this point the enthalpy ii=i,+O.SV? can be found. Now, knowing S; and ii, the remaining parameters, namely: T;, pb,etc. can be obtained from the same iS diagram or thermodynamic tables. The results of calculations will correspond to the given flight altitude. With a change in altitude the conditions of flowinteraction and hence the parameters at the stagnation point will also change. This relation is graphically reproduced in Fig. 4.7.4. The curves allow us to determine temperature IT;, and pressure pi as a function of velocity Vl and flight altitude H. The results of calculations of density with allowance for dissociation and ionization at the stagnation point for air as well as pure oxygen and nitrogen are set out in Fig. 4.7.5 [52]. The following conclusions can be derived from the analysis of these results: At M,= 18, when oxygen is already considerably dissociated, the value of Alp, reaches its maximum value. With an increase in M, the gas becomes completely dissociated and the density decreases. Then, with further rise in M,, ionization of .oxygen sets in, leading to a rise in specific heats and hence to some increase in density. The effect of variation of P;/L specific heats on the change in density of nitrogen is observed 15 only at very high Mach numbers M, when dissociation and 13 ionization take place. These processes occur not one after the If other as in the case of oxygen 9 but practically simultaneously. This is confirmed by the small difference of energies for the dissociation and ionization of Fig. 4.7.5. Density at point of complete nitrogen. Due to this the curve stagnation for nitfogen, oxygen and air of for nitrogen is more (p,=0.01 atm, Tw= 2 0 K . 9) monotonous than for oxygen. At Mach numbers M,< 13 basic dissociation of oxygen takes place and the curve of density of air is close to the corresponding curve for 0 2 . At M > 13, when dissociation of nitrogen begins to play a part, the ratio of , pi/p,for air will be similar to that for nitrogen because it then becomes the active component in air. Calculations of gas parameters behind a shock with allowance for variations in specific heats are discussed in more detail in book [lo].
13 17
7 v 9
21
25
M,
169
The parameters behind a shock can be calculated in a comparatively simple way using the general system of equations of an oblique shock, relations for determining the thermodynamic functions and the degree of dissociation of a heated diatomic pure gas. For this an equation of energy (4.2.6) can be used. Together with relation (1.5.13') for enthalpy it can be written in the following form:
Here the pressure and density are related to the characteristic parameters pd and pd respectively and the velocity is related to the characteristic quantity Vd. Assume that the degree of dissociation up to the shock is low and that the shock is very strong (pressure, density and temperature behind the shock considerably exceed their values before the shock). In this case, neglecting the first two terms on the left side of the equation (4.8.1), we get
Now we use the equation of momentum (4.2.4"). For very high velocities at which p l g p 2 it is written as
PI V
~ I ~
~ ~ 2 2c 2 f .
(4.8.3)
=p2Vnz,
we get
V n l = ( P z . / P z ~ ~ z )Vn29
P ~ I P ~ =V$ AVn (1 AVn). Introduction of the characteristic parameters pd, pd and results in
Vd
Substituting the given value in (4.8.2'), we write the following quadratic equation of the unknown variable ATn:
170
AERODYNAMICS
where
2 D=(7+a2) 9
(1+a2) az.
Equation (4.8.5) is solved simultaneously with equation (1.5.14). which can be rewritten in the form Further, (4.8.5) can be solved simultaneously with the equation of enthalpy (4.2.16), which under the condition of il< i2 takes the form
Here & represents an enthalpy with respect to the characteristic energy of is the velocity in terms of characteristic speed Vd. dissociation ud and The calculations are carried out by successive iterations. Strictly speaking this method of calculation for shock in relation to a dissociating diatomic gas can be applied to the calculation of the parameters of air behind the shock within a given approximation. For this the characteristic parameters of dissociation must be found for air as a model gas consisting of a mixture of nitrogen and oxygen in proportion to their mass content. As shown by the calculations, the parameters of air for real air differ from those obtained by the method devised for the diatomic model. Thus, for speed Vt11=1.5Vd= 8.1 kmlsec the density of real air behind the shock is p2= 14.7~1, which is higher by about 5% than the diatomic model.
9. Relaxation Phenomena
In the previous section we examined the methods of calculating gas parameters behind a shock with allowance for physicochemical changes and the condition of equilibrium of thermodynamic processes taking place in a gas. However, in the most general case these processes are characterized by their unbalanced nature, which seems to have some effect on the gas flow behind the shock.
9.1 Concept of nonequilibrium flows As is known from any course in thermodynamics, the assumption of thermodynamic equilibrium is related to the parameters characterizing the state of a gas in accordance with the mode of internal degrees of freedom. For example, at not very high temperatures comparatively (not at high velocities) an equilibrium between temperature and vibrational degree of
171
freedom is established which corresponds to the equilibrium between temperature and specific heat. At high temperatures (at high speeds), i.e. when gas dissociates, the equilibrium state is reached in the following way: With the development of dissociation the probability of triple collisions rises because the number of gas particles increases. This leads to recombination and a low rate of dissociation. There exists a moment of time at some temperature when the rates of direct and reverse reactions are equal and the gas enters a state of equilibrium which is characterized by a constant composition relation between the degree of dissociation from one side and temperature and pressure from the other. Finally, at still higher temperatures (at very high speeds) one can talk about the equilibrium processes of existing electronic states and ionization. In an equilibrium flow the internal degrees of freedom are abruptly established with an instantaneous variation of temperature, whereby the dissociation and ionization can be considered as the appearance of new degrees of freedom. In these cases there is no delay in establishing degrees of freedom, i.e. the time to achieve equilibrium is equal to zero. In practice the equilibrium flow is observed at supersonic speeds of flow over a body at Mach number M,=4 to 5 in conditions corresponding to altitudes of 10 to 15 km and less. This is explained by the fact that at maximum temperatures of the order of 1,0001,500K arising under these conditions the main part of the internal energy comes from the translational and rotational degrees of freedom which, during sudden changes in temperature, are established almost instantaneously. This is because it is enough to have only a few molecular collisions to achieve equilibrium. For this reason the translational and rotational modes of freedom are usually known as "active" modes. With a rise in speed and hence in temperature the significant part of the internal energy goes to vibrations and then to dissociation, excitation of electronic levels of energy and ionization. Practically speaking these processes are such that the energy modes are set more slowly than the translational and rotational modes because they need a fairly high number of collisions. Therefore the vibrational and dissociating modes are sometimes called the "inert" modes. Thus, inert modes are characterized by a delay in achieving equilibrium, known as relaxation. The time when equilibrium is reached in the flow, i.e. when the correlation is established between temperature and energy level, represents in itself the relaxation time. According to this discussion, the relaxation time serves as a criterion of the rate of diminution of the deviation of a gas from its equilibrium position, which is represented in a general case in the form of variation of energy distribution by,different modes of freedom. The following example is an intuitive representation of the relaxation process and relaxation time: Let us assume that two thermodynamically
closed; balanced gas systems are brought into contact with heat. As a result their equilibrium is disturbed. However, after heat exchange over a period of time both systems reach another thermodynamic equilibrium state. The process of transfer to the equilibrium state is known as the relaxation and the time required to establish equilibrium is called the relaxation time. The relaxation processes are determined according to which mode of freedom comes to excitation. If a sudden change in temperature results in vibrations the corresponding disequilibrium process is called the vibrational relaxation. It is characterized by sluggishness in the variation of specific heats with change in temperature. If the temperature rises the specific heat increases as a result of vibrations of atoms in molecules. The time during which the vibrational motion returns to equilibrium is known as the vibrational time of relaxation. In an unbalanced dissociating gas, for any instantaneous variation in temperature there results a sluggishness in the change of degree of dissociation. This phenomenon is known as dissociational relaxation. In connection with the difference in rates of formation of atoms and loss of atoms (rate of dissociation higher than recombination) a gradual rise in degrees of dissociation occurs. A balanced state of degree of dissociation is reached when the rates of direct and reverse reactions are the same. The time required to achieve a balanced degree of dissociation is called the dissociational time of relaxation. At temperatures up to roughly 10,OOOK the vibrational and dissociational relaxation processes happen to be the basic processes. The relaxation phenomena connected with the excitation of electronic levels of molecules and atoms and also ionization can be neglected because only a small part of the internal energy goes to these modes at these temperatures. An unbalanced state seems to have a considerable effect on different processes associated with gas flows at high speeds. In particular, the vibrational and dissociational relaxations change the gas parameters during transition through shock waves and during the flowinteraction of bodies. In its turn.this affects the process of friction, heat transfer and redistribution of pressure.
9.2 Equation of rate of chemical reactions Study of an unbalanced flow must include investigation of the motion of the medium and chemical processes which occur at finite rates. Formally speaking, this is characterized by addition of the equation for rates of chemical reactions to the usual system of gas dynamic equations. To simplify the investigation we use the equation describing a simple binary reaction of dissociation and recombination of pure dissociating diatomic gas which is represented by the general relation (1.5.1). In this relation the rates of dissociation and recombination are not equal and hence the chemical reaction is characterized by some difference in these reaction rates.
173
Equilibrium is,reached when the rate of formation of new molecules as a result of recombination of atoms is equal to the rate of loss of molecules dissociating into atoms. Thus the real rate of reaction of dissociation will be
where a is the instantaneous value of degree of dissociation. To find the general expressions for rates of dissociation and recombination it is necessary to use the relations of chemical kinetics. The recombination of two atoms A in molecule A2 during collision with a third particle B is expressed by the formula A +A +B+A2 +B, where B is a particle carrying
k~
away the energy of recombination; k~ is the constant of rate of recombination. According to this formula the rate of change of concentration of atoms during recombination is
Id*]}
[A]=mA'
=2
k [A]' [B]. ~
The square brackets indicate concentration in units of mol/cm3; k~ has the dimension of cm6/mo12. sec. For a binary mixture of molecules and atoms the sum of concentrations C M + C A = 1 ; consequently
[A2]==
(4.9.3) 2 m ~'
where m ~ma are the molecular weight of diatomic molecules and atomic , gas ( m ~ = 2 m ~concentrations C A and c~ are related to the degree of dis); sociation by formulas c~=a, C M = 1a. For concentration and density we also have the relations
Introducing relations (4.9.3) in (4.9.2), we get the relation for rate of recombination: r~ = ( d a l d t )=k~ ( p / m ~ )a2 (1 +a). ~ *
(4.9.4)
of mutual collisions. According to this scheme of reactions, the rate of dissociation will be
Here kDis the constant for rate of dissociation, B is the second particle participating in a collision with the molecules and providing it with the energy of dissociation. Introducing values of [A21 and [B] from (4.9.3) in (4.9.2'), we get
~ where Kp= k ~ / k for; balanced processes this quantity is defined as a constant of thermodynamic equilibrium. Under certain conditions this constant is valid even for unbalanced reactions. Keeping this in view and assuming da/dt=O in equation (4.9.6), we get the value of K,=zP~:/(I  a 3 in a
ma
balanced condition, where a, is the degree of dissociation in equilibrium condition. Recalling equation (1.5.2), we will get
 
c=(k,/m;)
( 1 a) Pd.
(4.9.8)
As can be seen from the expression (4.9.7), it is important to know the constants of the rate of recombination for investigation of unbalanced flows. This is enough to determine the rate of dissociation because it is possible to use this as a constant of thermodynamic equilibrium. The mean empirical values of the constant for recombination in air obtained at a pressure r.:o!x of about one atmosphere are show11 in Fig. Fig. 4.9.1. Variation in coefficient of rate of recombination 4.9.1. It is assumed that these values correspond to the lower limit of the rate of recombination. in air with temperature.
\ \ 1:
175
During the experiments with oxygen and nitrogen the measured values of constants of rates of recombination were of the order of 10151016 cm6/mo12sec. = (Wcllem)~ occurring in Weightwise the rate of formation of atoms the equation of diffusion (3.2.5) is connected with (4.9.7) by the obvious relation
9.3 Time of relaxation Equation (4.9.7) can be modified to some extent. For this purpose we use the concept of the constant of equilibrium which is useful under the conditions of unbalanced and balanced flows. Then the ratio e'IF/p can be replaced by the quantity az/(l a,) which, after substitution of (4.9.7), gives
(4.9.10')
Integration of this equation leads to the relation a (t), which characterizes the change in degree of dissociation as a func'' SeC D tion of time t beginning with its value a at time t = O and ending with the value at 10 equilibrium a =a,, which is reached in some time At (time of relaxation). To the order of this value the derivative daldt can be expressed in the form da/dta(a, a)/At. Consequently, according to (4.9.10') we have At~to. In this way the parameter to (4.9.10) can be approximately defined as the time of relaxation. The most important problem of the physics of relaxation processes is included in the determination of this parameter. Some data on the relaxation time
r 2 3 4 5 T.IO?~K Fig. 4.9.2. Experimental curves ( 1 s 2) for time of vibrational relaxation and theoretical curves (3, 4) for time of relaxation for dissociation of oxygen and nitrogen respectively at atmospheric pressure.
for,oxygen and nitrogen are obtained experimentally in shock tubes (Fig. 4.9.2). Analysis shows that the time of relaxation for dissociation is less than that for recombination by approximately the first order. Therefore to estimate the effects of relaxation it is desirable to use the time of recombination even though a rough estimate can also be made from the time of direct reaction. The same thing can be said of vibrational relaxation, keeping in mind that in an actual process the time required to excite vibrations up to a given temperature is less than the time taken to damp the vibrations in the opposite process beginning from that temperature. To estimate the relaxation time of oxygen or nitrogen the following approximate expression can be used:
t ~ = 7 . 7x lo'
a ( 1 1 . 7 ~ )(3)'12 0 T (PK)~
(4.9.11)
where index "K" indicates final values of parameters. Formula (4.9.11) shows that reactions take place at a faster rate if the initial density and temperature increase; this reduces the relaxation time. The same formula indicates that time to decreases with a rise in pressure, which is confirmed by the graphs in Fig. 4.9.3. It can be seen that an increase in pressure to the third order gives a tenfold reduction in relaxation time (pressure p in graph is related to atmospheric pressure ps at sea level). Investigations proved that the relaxation time during dissociation of air is less than that of pure nitrogen or t D , Sec oxygen. Some information from the kinetic IO~ theory of gases: In the kinetic theory of gases the relaxation time is defined as the time during which the motion of a 10'~ molecule undergoing collisions with other molecules having an indefinite character must in general be independent of the previous history of the motion of the molecules. Obviously this time tr depends on the number of collisions n suffered by a molecule in unit time and hence it may be assumed that t, is 10'~ inversely proportional to n, i.e. 3 5 7 ~ T I o  ~ , ~ K
where coefficient a, depends on whether the molecule collides with the wall or with another molecule. Besides, its value changes with the nature of the interaction. For example, to change the energy of vibrational motion it may
tr =arln,
(4.9.12)
be necessary to, have somewhat bigger collisions than to change the energy of translational motion. If a mixture of gases is considered the coefficient ar will depend on whether molecules of one kind interact with those of the other kind. This coefficient will depend on the type of process 'of interactions between molecules, which is determined by the properties of the gas medium and the material of the wall with which the molecules collide. The average number of collisions suffered by a molecule per unit time depends on the cross section of molecule A, average speed and number of molecules N in unit volume. The proposed number of collisions will be expressed by number of molecules n situated in volume A representing a ; "corridor" of its own kind formed by a molecule in unit time. Thus n= N A ~ From this it is possible to find the interval between two collisions of one molecule with another, i.e. t = I/(NA;). Accordingly the time of relaxation (4.912) will be tr = apt= ar/NAc. (4.9.13)
The average distance traversed by a molecule during this time indicates the relaxation path
r For estimation of the order of quantities tr and I the data on the properties of air at normal temperature and pressure can be used: N=2.69 x 1019 ~rn~; ;=4.5 x lo4cm/sec; A= lo4 cm2. Introducing these data into (4.9.13) and (4.9.14), we get &=ar sec; Ir=4ar lod5cm.
Experiments showed that for active modes of freedom (translational and vibrational) the coefficient ar will be of the order of unity and for inert modes (vibrational, dissociational and other internal degrees of freedom) the quantity a, can be high, approaching lo6. As can be seen from (4.9.13) the time of relaxation diminishes with an increase in density. Its relation to temperature will be more complicated because not only average molecular speed; but also ar and A depend on temperature.
9.4 Equilibrium processes Equilibrium flows have been studied in more detail than unbalanced flows from both the quantitative and qualitative points of view. The regions of flow in which equilibrium is established differ due to the unequal times of relaxation for excitation levels. As mentioned earlier, the time to establish an equilibrium in vibrational modes of freedom is a few orders greater than that for translational and rotational modes. Equilibrium of the gas composition mixture during its dissociation and ionization takes still longer.
Accordingly the scheme of an unbalanced process is such that achievement of equilibrium of one degree of freedom may be accompanied by the beginning of the relaxation process of another. In the more general case we encounter overlapping of the regions where equilibrium is established. An approximate model of the process can be presented on the basis of the principle of "freezing." It is assumed that the region where equilibrium is achieved is related to one or several degrees of freedom at a time when other modes are not excited. The sequence of these regions of equilibrium can be expressed by degrees of freedom with the rise in temperature in the following order: translational, rotational and vibrational modes, dissociation, excitation of electronic levels, ionization. For example, in considering the process of acquiring equilibrium in the vibrational mode the first two degrees of freedom can be assumed as completely excited. This process continues under conditions of "frozen" dissociation and ionization. Such a scheme is not acceptable for some gases, in particular for nitrogen at high temperatures, as ionization starts before the completion of dissociation. This is explained by the fact that the energies of dissociation and ionization in nitrogen differ between themselves by at least one and onehalf times. In this case the regions where equilibrium is achieved overlap. A similar phenomenon is observed in air at comparatively low temperatures. At temperatures above 3,200Kthe relaxation time for dissociation of oxygen is less than the time it takes nitrogen to achieve the vibrational mode. Hence equilibrium of dissociation in oxygen will be reached before vibrations start in nitrogen. Investigation of the flow of an unbalanced gas over a body is simplified if the characteristic time for this process is considerably less than the relaxation time of one of the inert processes and we get the conditions of "freezing" of the flow without the participation of this inert process. In particular, if the time of interaction of some part of the surface is less than the time required for chemical equilibrium but equal to the time of vibrational relaxation then on this part of the surface it is possible to consider the process with 'cfrozen" dissociation and ionization.
9.5 Effects of relaxation in shock waves During transition of a gas through a shock wave a part of the kinetic energy is converted into the energy of active and inert degrees of freedom where the equilibrium for active degrees of freedom, translational and rotational, is established in a time equal to time of relaxation t4, which is the same as the transition time for gas passing through the thickness of the shock. As this time is very small it may be assumed in practice that the active degrees of freedom are established instantaneously. So just behind the shock the temperature T; will be such that the specific heats of the gas remain constant.
According to*thisscheme the inert degrees of freedom immediately behind the shock are not excited. Since these modes have a finite relaxation time t: significantly higher than t,A and the transition time through the thickness of the actual shock the temperature will decrease until the inert modes (first vibrations and then dissociation, excitation of electronized levels and ionization) no longer reach equilibrium. In Fig. 4.9.4, which schematically illustrates the relaxation process, the time of relaxation t,A refers to temperature T i calculated on the basis of the beginning of motion on the leading surface of a shock wave where the temperature of the undisturbed gas is T,. Density p; corresponding to temperature T; is shown in this figure. Temperature T,< Ti corresponds to the equilibrium condition. The unbalanced gas flow process behind the shock is During accompanied by a rise in density to its equilibrium value pz > this process a small rise in pressure by comparison with the ideal gas flow is observed. Simultaneously the degree of dissociation (and at very high temperatures the degree of ionization) increases from zero to its equilibrium value.
Fig. 4.9.4. Relaxation process in shock wave (hatched region determines thickness of shock wave):
1 and 2front and rear surfaces of shock wave respectively.
The study of an unbalanced flow behind a shock consists in determination of the extent of the disequilibrium zone or the length of the relaxation path and the estimation of unbalanced flow parameters. Let us take the disequilibrium motion of a model diatomic gas along a stream line behind a strong normal shock. The system of equations describing this motion includes the equation of momentum (4.8.3) which can be written as

 
p , V : = p + p V2.
(4.9.15)
180 AERODYNAMICS
The equation of state and the equation for the rate of chemical reactions are given by (1.5.4') and (4.9.7) respectively. In these expressionsp, p, T,v represent nondimensional values of pressure, density, temperature and velocity at an arbitrary point on a stream line in the unbalanced zone. These quantities are obtained with respect to the corresponding characteristic parameters. The quantity a determines the unbalanced degree of dissociation at this point. The parameter C, appearing in (4.9.7), is calculated by formula (4.9.8). If the derivative daldt is expressed in the form
The determination of unbalanced parameters deals with the solution of this differential equation along with the other equations of the system. During integration the initial conditions are determined by the parameters immediately behind the shock wave which can be found on the assumption that dissociation is absent, i.e. at x=O the quantity a =O. At the end of the relaxation path these parameters reach equilibrium values corresponding to degree of dissociation a=a, at the equilibrium condition. Here for simplification the pressure in the relaxation zone can be taken as constant and equal to its unbalanced value just behind the shock. For this purpose we can apply the equation for rate of chemical reactions.
.
0.30 0.20
shock wave
2
1
LO^ X
Fig. 4.9.5. Variation in unbalanced degree of dissociation and density of oxygen behind shock wave (M1=13, Tl=2C)50K, p1=5.65 mm mercury column).
The results of numerical integration of equation (4.9.18) for an unbalanced flow of oxygen behind a normal shock wave are graphically shown in Fig. 4.9.5. The calculations were carried out on the basis of the experimental coefficient of rate of recombination kR = 8.4 1014 cm6/ mo12.sec. The curve on this figure helps .in estimating the unbalanced values of the degree of dissociation in
18 1
the zone of relairation for the case of an equilibrium in the vibrational mode. According to the nature of the variation in the degree of dissociation the density and temperature change from their respective values for a nondissociating gas to the values of balanced dissociation. For an approximate evaluation of the effect of disequilibrium on an airflow the same equations may be used with the diatomic model of air consisting of the additive mixture of oxygen and nitrogen. Here the parameters C in equation (4.9.17) can be determined for oxygen because the coefficient of the rate of recombination k~ is known more accurately from the experimental data for this gas. All the remaining parameters, particularly the degree of balanced dissociation a,, characteristic density, pressure, etc. are found for the diatomic model of air. The unbalanced parameters determined by the above method correspond to the condition under which the gas medium in the initial (undissociated) state has the excited vibrational modes and for such gas k =c,/c, = 1.33. If the equation of energy in the form i+V2/2=io is applied in place of equation (4.9.16) the initial condition of the heated gas in the first stagecan be avoided. Then the above unbalanced parameters in the zone of relaxation can be investigated depending on whether vibrational levels are initially excited or not. The results of such investigations for the air mixture of oxygen and nitrogen are reproduced in Fig. 4.9.6. The thick curves were P2,4, .,O J, obtained on the basis of instantaneous vibrational exciI0 tation and the broken curves were obtained in the absence of vibration. From these 8 results it follows that the 6 vibrations have significant value in the immediate vicinity of the shock. For 4 example, without consideration of vibrations the tempe Fig. 4.9.6. Effect of unbalanced dissociation on density and temperature behind shock wave (MI= 14.2, rature behind the shock is TI =300K, 1 mm mercury column). p,= equal to 12,000K while for the completely excited state it is about 9,800K, i.e. considerably lower. At the end of the relaxation zone vibrational excitation has practically no significance. Therefore in calculations of balanced dissociation it is possible to assume that the velocity for vibrational excitation is infinitely high and hence that the gas before the beginning of dissociation as fully excited. In Fig. 4.9.6 it can be seen that the extent of the unbalanced zone is comparatively small: it is of the order of 8 to 10 mm.
5
METHOD OF CHARACTERISTICS
1. Equations for Velocity Potential and Stream Function
The method of characteristics plays an important role in aerodynamics. It helps us to analyze the disturbed flow of an ideal (inviscid) fluid. With the application of this method it is possible to determine correctly the contour of nozzles for supersonic wind tunnels and to find the parameters of supersonic flow of wing profiles and cones of a flight vehicle. The method of characteristics is worked out for various cases as for the solution of a system of equations in steady supersonic twodimensional (plane or threedimensional axisymmetric) rotational and irrotational gas flows. Research connected with the application of the method of characteristics in calculations of a threedimensional flow over bodies has a wide range of applications. The present chapter deals with this method and its utility for the problems of supersonic twodimensional flows. The equations of the twodimensional plane steady motion of an inviscid gas can be obtained from (3.1.20), which under the condition of p= 0, aVA/at = aV,/at =0 takes the following form:
For twodimensional axisymmetric flows the equations of motion obtained from (3.1.36) under similar conditions (v =0, aVx/at=aVr/at=0) can be written in the form:
The equation of continuity for plane and axisymmetric flows having the form of (2.4.5) and (2.4.32) respectively can be written in the general form:
182
At E = 0 this equation coincides with the continuity equation for twodimensional plane motion in Cartesian coordinates x, y. If E = 1 we get the continuity equation for a twodimensional axisymmetric flow in cylindrical coordinates y(r), x. Accordingly the equations of motion (5.1.1) for both types of flow can be assumed to have been written in general form. Taking the partial derivative of the equation of continuity (5.1.3), we get
The partial derivative aplax can be replaced by the expression ap/ax= (dpldp) (dpldx), where dpldp = l/a2and aplax is found from (5.1 .l) in the form
Then
This equation is the basic differential equation of gas dynamics for a twodimensional (plane or axisymmetric) steady flow, which must be satisfied by the velocity components V,, V y . As this equation relates velocities it is known as the basic kinematic equation. If the flow is potential,
where
Equation (5.1.8) is the basic differential equation of gas dynamics for a diatomic potential steady flow and is known as the equation for velocity potential. Thus unlike (5.1.7) this equation is applied only in steady irrotatiocal gas flows. If a twodimensional gas flow is rotational it is necessary to use the stream function y for this flow. The velocity components, expressed through function y, have the following form according to equation (2.5.5):
Replacing p by formula (3.6.3 1) in which stagnation density po along the given stream line has constant value, expressions (5.1.9) can be written in the form:
where Y= VV. /, The calculation of a rotational gas flow lies in solving the differential equation for stream function y. To get this equation differentiate V and V, in (5.1.9') with respect to y and x respectively: ,
Considering the expression (3.6.22) for the square of sound velocity and the equations (5.1.9'), the relations obtained for derivatives aV,/ay and aV,/ax can be written as following:
The derivations aV2/a.y and aV2/ax appearing here could be determined with the help of the relations (5.1.9'). Combining these relations, we get (ay/a~)~+ (at,~/ay)~v2y2 V2)2/(kl). = (1 (5.1.12)
Differentiating (5.1.12) with respect to x and y, the following relations can be obtained:
METHOD OF CHARACTERISTICS
18 5
Using equation (3.6.22) for the square of the sound velocity and the relations (5.1.9'), we can transform the above relations into the following:
.
52);
(5.1.13)
a2 + vy ,
axa~
av2(,  ~ 2 ) 1 / , ,  1 )
x 1   +  V2 (1  T/2)ll(kl).
2) ,fe
a~
(5.1.14)
The derivatives aV2/ay and aV2/ax can be determined from these relations. Substituting these values in (5.1.10) and (5.1.11) respectively, we get:
Subtracting the first equation from the second, we get the relation for vorticity:
Vorticity can also be determined with the help of equation (3.1.22'), which is obtained for the condition of steady flow of an ideal gas in the form grad (V2/2)+rot V x V= (l/p) grad p. Taking the gradient of the equation of energy (3.4.14), we get grad (V2/2)=  grad i. Consequently equation (5.1.16) can be rewritten in the form: rot V x V=grad i(llp) gradp. (5.1.17)

(5.1.16)
 
(5.1.18)
Rotational supersonic flows of a gas can be thermodynamically characterized by the variation in entropy from one stream line to the other. Therefore it is convenient to introduce a parameter into the calculation that can reflect the given variation in entropy as the peculiarity of rotational flows. From the second law of thermodynamics we have
TdS =di(dplp),
or in the vector form
(5.1.19) (5.1.20)
Let us find the vector product rot V x V, keeping in mind that according to (2.2.12) the vector
  
and hence the component (rot F)x = (rot the third order, we get
il, iz i3 r o t V x V = 0 0 (rot7)3 , V v." 0 X in which the component (rot  =  V)3 avy avx ax
a~
This gives
METHOD OF CHARACTERISTICS
187
Accordingly we get the relation for the projection of vector rot V x V along the normal to a given stream line as:
e
( r o t V x ~ ~ = ( r o t ~ x ~ ~ + (v):= v2 vx rot
(%%)2.
(5.1.22)
From (5.1.20) it follows that this component of the vector can also be expressed in the form (rot V x V),
 
= T (dS/dn)
As a2=kRT we obtain from this and (3.6.22) the expression for temperature:
Substituting vortex strength from this in (5.1.15) we get a differential equation for stream function:
2. Cauchy's Problem Equations (5.1.8) and (5.1.25) for velocity potential and stream function respectively are nonlinear differential equations of the second order. Besides terms with secondorder partial derivatives these equations also contain free terms without partial derivatives. Therefore they are not homogenous. The solutions of equations rp = rp (x, y) and ty = ty (x, y) are geometrically represented by integral surfaces in space which can be determined by systems of coordinates x, y, rp or x, y, ty: In these systems the plane x, y is regarded as
188
AERODYNAMICS
the main plane. It is known as the physical plane or plane of independent variables. The process of finding solutions satisfying given additional conditions on a curve in the neighborhood of some given (initial) curve y = y (x) forms the content of Cauchy's problem. The values of the functions q (x, y) or y (x, y) being investigated and their first derivatives qx(tyx) or p,(ty,), known as Cauchy's initial conditions, are examples of such additional conditions. It should be noted here that a given function that is itself on the initial curve, for example q [x, y (x)], and one of its first derivatives q, =aq/ay automatically determine the other derivative qx= aqlax. This follows from the relation
which is obtained from the formula for total derivatives of the complicated function of two variables x and y on the assumption that the initial curve is given and hence the relation between y and x is known. From the geometrical point of view Cauchy's problem consists of finding an integral surface in the space x, y, q (or x, y, y) passing through some given curve in space. The projection of this curve on the x, y plane also represents an initial curve y = y (x) on this plane. The solution of Cauchy's problem is applicable to the study of supersonic gas flows and the credit for developing the corresponding method of characteristics belongs to the Soviet scientist Prof. F. Frankly. To study Cauchy's problem we express equations (5.1.8) and (5.1.25) in a general form:
0
Fig. 5.2.1. Initial curve AB
On
which desired function and its first derivatives with respect to and y are known.
where u= q (or yxx) s= q, (or yxy), t = , q,, (or yyy)are the secondorder partial derivatives; the values of A, B and C are equal to the coefficients for the corresponding secondorder partial derivatives; the quantity H is determined from the free terms in equations (5.1.8), (5.1.25). ~ eus find the solution of equation (5.2.1) t in the neighborhood of the initial curve AB (Fig. 5.2.1), which is in the form of a series. At Some point M (XO, the function being YO) investigated will be
where p (x, y) is the value of this function at a given point N (x, y) on the initial curve; Ax=sox, Ay= yo y. In place of p in (5.2.2) the function y can be used. The series (5.2.2) gives the function we are looking for if the values of function a, (or y) and its partial derivatives of any order anp/axn,plp/axnl, etc. exist and are known on the given curve. Since on this curve the first derivatives [denoted by p= q, (or yn) and q= p, (or yy)] are given it is necessary to find second derivatives and the derivatives of higher order on this curve. In this way the solution of Cauchy's problem is connected with the investigation of conditions in which it is possible to determine actual derivatives on a given curve. Here we confine ourselves to the determination of secondorder derivatives. As there are three such derivatives (u, s and t) only three independent equations can be formulated. The first of them will be equation (5.2.2), which is satisfied on the initial curve AB. The remaining two equations can be obtained from the following known relations for the complete differentials of the functions of two independent variables existing on this curve:
Hence the system of equations for determining second derivatives will be written in the form:
This system of equations is solved for the unknowns u, s, t with the help of determinants. If the main and the partial determinants are expressed by the notations A and AU, As and At, respectively, From these relations it follows that if the main determinant is not equal to zero on the initial curve AB the secondorder derivatives u, s and t can have unique values. Let us consider a case where the curve is selected in such a way that the main determinant lying on it is equal to zero, i.e.
This gives
From the course in mathematics we know that under the condition when main determinant A of the system of equations (5.2.3) is zero, i.e. on' the curve expressed by the equation (5.2.5), the secondorder derivatives u, s, t (5.2.4) are either multivalued or in general cannot be determined in terms of p, p and q. Let us take quadratic equation (5.2.5). Solving it for derivative dyldx, we get
These equalities determine the inclination of the tangent at every point of the initial curve on which the main determinant A = 0 . It can easily be seen that the equations (5.2.6) are the differential equations of two families of real curves if B2AC >0 . These curves, at each point of which the main determinant of system (5.2.3) is equal to zero, are known as the characteristics and equation (5.2.6) is called the characteristic equation. From the above discussion we get the condition under which it is possible to find a unique solution for secondorder derivatives on the initial curve: An elemental arc of this curve must not coincide with the characteristics. Similarly while determining higherorder derivatives uniquely the same conditions A#O is to be satisfied. Hence if A#O, all the coefficients of series (5.2.2) can be determined uniquely from the data on the initial curves. In this way the condition A#O is the necessary and sufficient condition for solving Cauchy's problem. The problem is very important in the mathematical theory of partial differential equations and formula (5.2.2), generally speaking, can be used for calculation of a gas flow. However, from the point of view of physical principles and in particular the calculation of supersonic gas flows the problem of determination of a solution from the data on characteristics, i.e. the method of characteristics, is very important. This method can be obtained from the analysis of Cauchy's problem and may be described as follows: Let us assume that the initial curve AB coincides with one of the characteristics and that not only the main determinant of system (5.2.3) but also the auxiliary determinants are zero, i.e. A, =A, = A t =0 along it. If, for example, determinants A and At are equal to zero the remaining determinants will be automatically zero. To prove this let us find the auxiliary determinants:
METHOD OF CHARACTERISTICS
191
A 2B A, = #x dy 0 dx
H dp dq
2Bq' HI,
(5.2.9)
where p' =dp/dx, q' =dqldx. As dxf 0 the condition A t =O leads to the equality A (y'q' p') 2Bq' H = 0.
(5.2.10)
From the condition of A = O we get equation (5.2.5). Multiplying this equation by (p'y'q') and equation (5.2.10) by yt2and then adding, we have H~'~2By'p'+ C(p'yPq')=O. Comparing this expression with (5.2.7) it can be seen that A,,=O. Now if equation (5.2.5) is multiplied by q' and equation (5.2.10) by y' then, after subtracting the products, we get Ay'p'
+Hy' + Cq' =0
which, according to (5.2.8), corresponds to A,. The condition of zero values of all the determinants, as shown in the theory of the system of algebraic equations, signifies that solutions of system (5.2.3), although not singlevalued, do exist. Here if one of the solutions, for example of u, is the final solution then the solutions for s and t will also be final.
3. Characteristics
3.1 Conditions of existence Equations (5.2.5) and (5.2.10) determining the conditions under which solutions for u, s and t exist, even if multivalued, are known as the conditions of existence. Geometrically speaking the first of these equations represents two families of curves of characteristics in the physical plane x, y; the second expresses two families of curves of characteristics in the plane p, q. The characteristics of different families in these planes are known as conjugate families. The solution of a problem concerning a supersonic flow of gas found from the solution of characteristic equations (or of conditions of coexistence) also happens to be the solution of a basic equation of gas dynamics, (5.1.8) or (5.1.25). Its proof follows from the socalled theorem of equivalence according to which the characteristic equations (5.2.5) and (5.2.10) are equivalent to the basic equation (5.1.8) or (5.1.25) (the proof of this theorem is given in [21]). From the geometrical point of view the proven equivalence indicates that the solution of equations of characteristics also means transformation of some plane x, y to the plane p, q. That is, the points on the curves determined
by differential equation (5.1.8) correspond to the points on the curves determined by differential equation (5.1.25). In this way at each point on the characteristic in plane x, y there exists a point on the characteristic in plane p, q. Obviously the given correspondence can be established by various methods depending on the given graphical conditions. As will be seen further on, this relation particularly facilitates the use of characteristics in calculations of gas flows. From the above discussion it follows that the distinct property of characteristics lies in the fact that the initial conditions cannot be given arbitrarily along characteristics but can be given along a curve that is different from the characteristics.
3.2 Existence of characteristics Type of characteristics: From (5.2.6) it follows that the roots of quadratic characteristic equations (5.2.5) can be real (equal or unequal values) and also complex conjugate roots. The difference in roots is determined by the expression B2AC=6. In a case where 6 >0, the equation (5.2.6) gives two different families of real characteristics. The quantity 6 = 0 determines two equal roots which correspond to two coinciding characteristics, i.e. factually to only one characteristic. Finally, if 6 < 0, the roots of the equation represent a pair of imaginary characteristics. As the roots of characteristic equations depend on coefficients A, B, C of differential equation (5.2.1) it is the convention to classify these equations according to the type of characteristics. If 6 >0 equation (5.2.1) will be of the hyperbolic type, if 6 = 0 it will be of the parabolic type and for 6 < 0 it will be of the elliptic type. In equations (5.1.8) and (5.1.25) for velocity potential and stream function respectively the coefficients A, B and C are determined identically, namely,
(5.3.1)
(5.3.2)
: .: where the total velocity V= 4 V + V Thus for regions of gas flow with supersonic speeds (V >a) the equations will be of the hyperbolic type and for regions of flow with subsonic speeds (V< a) they will be of the elliptic type. At the boundary of these regions the speed is sonic (V=a) and the equations will be of the parabolic type. Characteristics in physical plane: The characteristics in plane x, y are determined from the solution of differential equation (5.2.6) in which the plus sign corresponds to the characteristics of the first family and the minus sign to the characteristics of the second. The quantity I1 =dy/dx which can be directly calculated from the expres
METHOD OF CHARACTERISTICS
193.
sion (5.2.6) and the plus sign in this expression determines the angular coefficient of the characteristic of the first family, while &=dy/dx with the minus sign determines the angular coefficient of the characteristic of the second family. Recalling expressions (5.3.1) and (5.3.2) we get the equation for characteristics in the x, y plane:
The characteristics in a physical plane have a particular physical meaning which can be established if angle , ~ lbetween the velocity vector 7 at s ~ m e point in the flow (Fig. 5.3.1) and the direction of characteristics at that point is determined. This angle is determined with the help of equation (5.3.3) if it is transformed to the local system of coordinates XI, yl with origin at point A and axis XI coinciding with the direction of velocity vector With this choice of axes of coordinates V,= V, V,=O and consequently,
r.
+ (M2 1)'1'.
From this it can be seen that jr represents by itself the Mach angle. Thus a very important property of characteristics is established. It indicates that at every point on the characteristic the angle between the tangent to it and the velocity vector at this point is equal to the Mach angle. Consequently the characteristic by itself represents a line of weak disturbances (or Mach line) having the form of a curve in the general case. 0 Determination of characteristic as the i . 5.3.1. Scheme ~~~h line is directly to the two F gsical meaning of representing P ~ Y characteristic: dimensional plane supersonic flow. If a Idirection of twodimensional space (axisymmetric) family at point characteristic of first A; 2characteristic supersonic flow is considered the Mach of fi,t family with angular coefilines (characteristics) may be regarded as cients a,; 3direction of characteristhe generators of a rotating surface which tic of second family at point A; 4is the slant Mach cone with its apex at the characteristic of second family wit11 angular coefficient A2. points of disturbance (on characteristics). The surface enveloping the region of disturbance is known as the wave surface or the space Mach wave. The compression waves arising in a gas where supersonic motion is characterized by a rise in pressure were discussed in Chapter 4. However, such motion can be accompanied by a decrease in pressure, i.e. the supersonic flow will be accompanied by expansion and the Mach ljnes will characterize expansion waves. The characteristics, which are generally curved lines (for a
194
AERODYNAMICS
plane flow) or surfaces formed by the rotation of these lines (for an axisymmetric flow), coincide with the corresponding Mach lines. If there are Mach lines (characteristics) in a flow in the form of straight lines they correspond to the plane expansion waves having the same direction of velocity of propagation. The Mach lines corresponding to expansion waves will be further known as lines of weak disturbances, using the accepted terminology for weak compression waves. Here it is necessary to bear in mind that in an expanding supersonic flow there cannot be any otherfhan weak expansion waves because otherwise the possibility could arise of formation of strong expansion waves (expansion shocks), which cannot occur in real flows. If the flow velocity and speed of sound are known at a given point in the physical plane the above property of characteristics helps us to determine the direction of characteristic at that point by calculating the Mach angle from the formula p= + sin' (l/M). The angular coefficientsof characteristics in relation to the coordinates x, y (see Fig. 5.3.1) will be obtained from the equation
8 where , is the angle of inclination of the velocity vector from the x axis; the plus and minus signs refer to the characteristics of the first and second family respectively. Equation (5.3.4) is the diffirential equation for the characteristics in the physical plane. Characteristics in plane p, q: If the quantity y' in equation (5.2.10) is replaced by first root (y'=Al) of the characteristic equation (5.2.5) the equation soobtained:
will represent the first family of characteristics in plane p, q. Similar substitution of y' by second root y1=12 gives the equation of the second family of characteristics in the same plane:
The equations of characteristics (5.3.5) and (5.3.6) can be transformed by the use of the property of the roots of quadratic equation (5.2.5), according to which Taking the first family of characteristics and introducing the relation Ah2B= 12A from (5.3.7) into (5.33, we write the equation A(1zq' +p') +H= 0. Similarly, for the characteristic of the second family we find A(A1q' +p') (5.3.8) (5.3.9)
+ H=O.
Taking into acmunt expression (5.3.4), equations (5.3.8) and (5.3.9) can be rewritten in the form
where the plus and minus signs correspond to the characteristics of first and second families respectively.
3.3 Property of orthogonal conjugate characteristics If the differentials in the equation for characteristics are replaced by finite differences the equations obtained will be the equations of straight lines in the corresponding planes x, y and p, q. Let us study the equations for conjugate characteristics, in particular for the characteristics of the first family in the x, y plane and of the second family in thep, q plane. From (5.3.4) it follows that for the element of a characteristic that is a straight segment in the x, y plane the equation has the form
where xo, yo are coordinates of some fixed point; 1 1 is the angular coefficient calculated on the basis of gas parameters at this point; x, y are variable coordinates. According to (5.3.9) the equation of the element of a characteristic of the second family in thep, q plane can be written in the following way: where po, qo are values of functions p and q at point xo, yo in the physical plane; the angular coefficient A1 and the values of A, H are worked out from the gas parameters at this point; p and q are variable coordinates. From these equations it can be seen that the inclination of a straight line in the y, x plane is determined by the angular coefficient I I and in the p, q plane by the angular coefficient 1/11. Similarly it can be shown that the element of a characteristic of the second family in the x, y plane has angular coefficient 1 2 and the element of a conjugate characteristic in the p, q plane has angular coefficient 1/12. From this it can be said that the conjugate characteristics in the two planes are mutually perpendicular. The above property facilitates the determination of the direction of characteristics in thep, q plane if the direction of the conjugate characteristics in the physical plane is known. Let us assume that the velocity components V,O, V and the values of functions PO,qo are known at some point P(xo, yo) , o in the x, y plane. The directions of Mach lines at this point (Fig. 5.3.2) can be determined by equation (5.3.4'). The element of a characteristic of the second family in the form of a straight line in the p, q plane given by equation (5.3.9') corresponds to the element of characteristic PN of the first family in
196
AERODYNAMICS
the,x, y plane. This straight line is perpendicular to line PN but does not pass through point P' with coordinates po, qo due to the independent term in equation (5.3.9'). Therefore to construct the element of a characteristic it is necessary to determine distance 6, from point P' to the element using the rules of analytic geometry. Similarly the conjugate characteristic of the first 2 family in the p, q plane which is perpendicular to line PM at distance 6 from point P' can be constructed (see Fig. 5.3.2).
The property of orthogonality of conjugate characteristics exists in the case of a potential flow for planes y, x and Vy,Vx (hodograph plane), and also in the case of rotational flow for which plane p, q is replaced by the same hodograph plane V,, V,.
3.4 Transformation of equations of characteristics in hodograph plane of velocity Let us transform equation (5.3.10) into a form such that it can determine the characteristics in the hodograph plane where the velocity components are the coordinates or independent variables. To do this, differentiate (5.1.9') with respect to x and calculate derivatives dpldx and dqldx:
METHOD OF CHARACTERISTICS
197
After substituting (5.3.1 1) through (5.3.13) in (5.3.10) and replacing the values of dy/dx by angular coefficients of characteristics A1,2 and the functions we get tan (p T p) by the corresponding values of
6= VY/V,ax,
V = V/Vm,,
V/a =M,
Recalling polar angle 8, the following expressions can be written for the projection of the velocity vector:
Differentiating these expressions with respect to x: dFx =c o s p  v s i n p dP ; dx dx dx dVy dfj sin /3+ v c o s P =dx dx dx
1 I
(5.3.16)
Inserting (5.3.15) and (5.3.16) in (5.1.14) and recalling that tan2,u=(M2 l)l, sin2 p =M2, we get
198
: , ik%
AERODYNAMICS
. sin p&,'
1 v2
Y
cos B
A1,2
k1 .
2k
RV
1  (1  P)
Taking into account that [(k  1)/2] (1  V2/VZm,) =a2/V2,, = V2sin2p, we get dV (A2,1 sin P +cos P) tan2 p dx +dP E A2,1 cos p  sin /3 Y sin2 p cos /3 dx tan2 p +k2 (~2.1 /3  sin p) (sin2 p cos2 /I) cos sin4 p dS . =o. (5.3.17) X (sin2 p  cos2 P) (12,1 cos p sin p) dn
Taking into account tan(p T p) and A1,2 = tan (p p) the separate expressions appearing in (5.3.17) can be transformed into the following form:
A2,1=
sin2 p cos p sin (P p) (&,I cos psin /3) (sin2pcos2 /l)=cos(P+p) sin2 p cos sin (P p) '[tan (pTp)cos psinp] cos @+p) cos (pp)=cosip?p)
A1,2
sin2 p cos /3 cos (p F p) sin (P p) sin p cos 8 sin /? cos j l . % sin p cos (B ,u) cos ( p  p)= cos (B k p) cos (B k p)' cos (B k ~ r ') (5.3.19) sin4 p (sin2 p  cos2 p) (A2,1 cos P  sin p) sin4 p cos (pi p) cos (p  11) [sin P  tan (P T p) cos p] sin3 p + sin4 p cos ( p T p) (5.3.20) cos (p+p) cos (pp) sin p = 'cos ( p + p ) '
Taking into account the expressions (5.3.18) through (5.3.20), the equations (5.3.17) for the characteristics of the first and second family respectively take the form:
METHOD OF CHARACTERISTICS
199
dV tan
dx sin p sin tan p dx sin3p dS ,udp& 0; (5.3.21) y cos ( j 3 f , ) u kR cos (P+LL) dndx sin p sin P tan p dx sin3p E + t a n pdpe   . cos(Pp) , dSSO, (5.3.22) V y cos(Pp) kR dn
. .
+ .
The gradieht of stagnation pressure dpb/dn can be used in the place of the gradient of entropy dS/dn. For this purpose we will use relations (4.3.6) and (4.3.20), which give the formula for the difference in entropy:
Since c, (k l)=R, the calculation of derivative with respect to n with the notation of dSz/dn = dS/dn, results in
In this way the equations of characteristics in the hodograph plane can be written as follows: dV dx sin 11 cot pdB&  . cos /3 sinp) v Y (P + kpi cos(P+p) dn dV dx sin p sin , u cot p jT+dpc Y cos (P P> dx sin2p cos , dph +. cos (/311) u =o. kpi dn Let us introduce the new variable
9
which indicates some angle. The ratio dVIV may be expressed in the form dI/I I 
( :*I
cot p = Consequently
4 ~ 1=~ 
(5.3.28)
200 AERODYNAMICS Thi~ integral can be divided into elementary integrals. For this purpose we use the substitution
Then
we get
k t tan 1 & s 2  t a n  1 $ s . s . max k1 A&i2 Replacing i in (5.3.30) by M from (5.3.28), we can write l
(5.3.30')
From equations (5.3.30') and (5.3.31) it can be seen that the angle CLI happens to be the function of 1 M) only and hence it can be calculated in advance. (or This simplifies the calculation of supersonic gas flows by the method of characteristics. The values of w for different Mach numbers M at k = 1.4 are given in Table 5.3.1. In the same table the angles of inclination of the line of disturbances obtained from the formula p=sin'(l/M) are set out. Using angle w in (5.3.21) and (5.3.22), we get the equations for characteristics: d(wTp)E
(5.3.32j
METHOD OF CHARACTERISTICS
20 1
TABLE 5.3.1
M
uO
iuO
u0
iu"
202
AERODYNAMICS
The corresponding substitution in (5.3.25) and (5.3.26) gives: dx sin b sin p  dx sin2, cos p u d(oTP)E. 17. . &=O. (5.3.33) Y cos(P+p) kp, cOs(P_+p) dn Equations (5.3.32) and (5.3.33) correspond to the most general case of a supersonic twodimensional (plane or axisymmetric) rotational (nonisentropic) gas flow.
3.5 Equations of characteristics in the hodograph plane for particular cases of gas motion The form of the equation of characteristics (5.2.5) in the physical plane is unique for all the cases of gas flow if this flow is supersonic and twodimensional. But in the hodograph plane the equations for characteristics will be different depending on the type of flows. If the twodimensional flow is irrotational then, by (5.1.23), the entropy at all points in space will be constant (dS/dn=O) and hence the equations of characteristics assume their most simple form:
dx sin p sin pFO, ~(coT~)E. Y cos(Ptiu)For a plane nonisentropic flow ( E =0) dx sin2p cos p . dS =o. d(mTB)' ~ ~ ~ C O S ( dn + ~ ) P In the most simple case of a plane irrotational flow (dS/dn =0) we have
Integrating (5.3.36), we get P = + otconst. Inserting some fixed values of angles PI acd p 2 in place of the constant, the first of which corresponds to the plus sign and the second to the minus sign, we find equations of characteristics in the form
Thus, unlike equation (5.2.5) for the characteristicsin the physical plane and equations (5.3.32), (5.3.34) or (5.3.35) for the characteristics in the hodograph plane having differential form, the corresponding equations (5.3.38) of the characteristics of a plane isentropic flow have a simple form. Geometrically. these equations determine two families of curvescharacteristics lying in a
METHOD OF CHARACTERISTICS
203
circular ring with an internal radius of I,= 1 and an external radius of I = [(k+ l)/(k 1)]1/2 (Fig. 5.3.3). The constant of integration and the plus sign before functions o(A) refer to the characteristic of the first family while the constant pz and minus sign before w(A) correspond to the characteristic of the second family. These curves are epicycloids which can be obtained by following the motion of the points on the circle of radius 1/2 (A,  1) (see Fig. 5.3.3). The polar coordinates of the points of the epicycloid happen to be angle of inclination jl of the and the Fig. 5.3.3. Epicycloidscharacteristics velocity I . of plane supersonic flow: Analyzing the graph on which Icharacteristic of first family; 2the network of epicycloids is reprecharacteristic of second family. sented, it can be said that the increase in velocity based on the expansion of flow corresponds to the higher values of angle 8, i.e. to the considerably larger deviation of flow from the original direction. At low velocities the deviation of flow direction will be small. The angle of inclination of the velocity vector is directly determined by angle w whose physical significance can be established from (5.3.37). Let us assume that the constants of integration P1,2=0. This indicates that expansion of flow begins from the conditions of ,3 =0 and M = 1. Accordingly the I quantity P= + w (M) will represent the angle of flow deviation during its isentropic expansion from the point where M = 1 up to the state characterized by some arbitrary Mach number M > 1 which is the upper limit of integration (5.3.27'). From the above discussion the difference between the angle of inclination is of flow at M > 1 from some original direction and the angle ,8= _+o obvious. Here jl determines the total turning of flow during its expansion from the condition of Mach number M = 1. The angle of deviation of flow at some arbitrary section can be determined in the following way: Let us assume that the initial Mach number M I > 1 is known. This may have increased during expansion of the flow to the value M2> M I . The Mach numbers MI and Mz correspond to the angles of flow inclination w l and m2 from the flow direction at the point where Mach number M= 1. These angles could be obtained graphically with the help of one of the epicycloids of the expressions (5.3.31) or Table 5.3.1. From the values of wl and w2.the angles of inclination of velocity vectors can be found.
204 AERODYNAMICS
Examining, in particular, the characteristic of the first family, we get Consequently the angle of inclination from the original direction will be The calculations may be carried out in reverse order, i.e. by determining the corresponding local Mach number M2 from the known angle of flow inclination A 8 and the initial Mach number MI. So we calculate from (5.3.39) (in the example given it is also desired to know the characteristic of the first family): From the graph or Table 5.3.1 we find the values of A2 or M2 corresponding. to the value of o 2 . I t is useful to calculate the critical angle of flow deviation. This is needed to obtain the maximum velocity Vmax. Let us assume that the flow begins to turn from the initial value of Mach number M = I . In this case the critical value of the angle of flow deviation could be found from formula (5.3.31), in which it is required to take the value of M for the value of Vmx equal to infinity:
or for k = 1 . 4 Thus a supersonic flow cannot turn by an angle exceeding w and theoretic, ally it can be said that a portion of space remains empty of gas. If the initial Mach number M at which the flow deviation starts is more than unity the angle of this deviation from the flow direction at M= 1 will be o (NI) and the critical angle of flow deviation corresponding to this Mach number will be
Pmax =wmax w
(M) =  (&~)W(M),
At hypersonic flows the calculation of function o and the respective angle of flow deviation becomes simple. Actually at M B 1 the terms appearing in (5.3.31) can be represented in the following form with an accuracy of higher order:
Then
All the above relations are obtained for a perfect gas. However, at very low pressure the properties of a perfect gas are not observed. Therefore these calculated angles of flow deviation are not realized in practice and have only theoretical importance. The formula corresponding to (5.3.39) takes the form
Determination of the parameters of a disturbed supersonic gas flow is related to the solution of a system of equations for an conjugate characteristics in the physical and hodograph planes if the initial conditions are in some respects given in the form of Cauchy's conditions. In the general case of a twodimensional nonisentropic flow this system takes the following forms: for characteristics of the first family
(5.4.1)
(P p);
(5.4.3)
where I, c, m, t are coefficients: sin2p cos p . sin p sin p l=cos (/?+p) ' C=cos (/?+p) '
sin B sin p sin2 p cos p. t= cos (p p)' cos (8 p) ' for a twodimensional isentropic flow
m=
Therefore the system of equations is simplified as follows: for characteristics of the first family
dy = dx tan (B +p);
d(o  8)s(dx/y) I= 0 ;
AYB= tan
~ B +,uB)AxB; B
(5.4.10)
where
IB =
CB =
tA=
sin2 ,ug cos PB. ' cos (PB+PB)' sin2 cos P A cos (PAPA)
(5.4.15)
Equations (5.4.10) through (5.4.13) are written on the assumption that the coefficients I, m, c and t remain constant along the elements of characteristics BC and AC and are equal to their values at the initial points B and A.
Fig. 5.4.1. Scheme for calculation of velocity at point of intersection of two characteristics of different families: aphysical plane; 5hodograph plane.
The variation over unit length of normal AS/An is calculated in the following way: From Fig. 5.4.1, a it can be seen that the distance from point B to point A is
Anm(AC) sin PA +(BC) sin pB,
where
Introducing the notation e = (xc XA) sin P A cos (FB PB), f =( X C XB)sin PB cos (FA P A ) ,
we get
208 AERODYNAMICS AS Sc =ASB +S B=(BC) sin ,UB +S B=( s A  S B ) f + ~ ~ . (5.4.17) An f+e The entropy gradient can be replaced by the gradient of stagnation pressure using equation (5.3.24):
The stagnation pressure at point C will be (5.4.19) f+e To determine the coordinate X C , yc of point C it is necessary to solve the system of equations (5.4. lo), (5.4.12) for the elements of characteristics in the physical plane: Y C ye= (XC X B )  tan (DBIPB); YC YA =( X C  XA)tan (PA PA). The graphical solution of these equations is shown in Fig. 5.4.1, a. From the value of x c obtained the differences AXB=xc X B , AXA=xc X A appearing in equations (5.4.11), (5.4.13) are determined. The increments AmB, d o A , A ~ and APA happen to be the unknowns in these equations. The number of B unknowns should be reduced to two according to the number of equations in the system. For this purpose let us write the obvious relations: AW~=W~CO~=~U)B+ OBWA; (5.4.20) A~A=~C~A=A~B+,~BPA. Equation (5.4.13) will be transformed with help of these equations into: Poc 
( P L  P ~ B ) ~ + ~ ; ~ .
Solving this equation together with (5.4.1 1) to find the variable Ape, we get
From the value of ADB SO obtained we calculate the increase in function from equation (5.4.1 1):
Now it is possible to find the angles for the point C: PC =Ape PB; wc =Awe we.
+ +
Now from the value of o c so obtained the Mach number Mc and Mach angle ,UC at a point C can be determined from Table 5.3.1. The scheme of the graphic solution of the system of characteristic equations in the hodograph plane determining angle PCand Mach number I c (Me) is shown in Fig. 5.4.1, b where B'C' and A'C' are the elements of characteristics of first and second families corresponding to the elements of conjugate characteristics BC and AC in the physical plane. From the value of Mach number MC (or Ac) obtained the parameters like pressure, density, temperature, etc. can be worked out if necessary. The accuracy of the calculated parameters can be increased if the values of angles P and Y between points A and C and points B and C are used in , equations (5.4.10) through (5.4.13) in place of le, m ~ CB, t~ as a first , approximation, i.e. if they are replaced by the respective values of:
The more accurate coordinates xc, yc of point C will be obtained from the equations: AYB = tan ( ~ g4 )p$') AxB; AYA=tan (By) ~ 2 )AxA. ) Second problem: This concerns the calculation of velocity at a point of intersection of the solid wall and the characteristic. Let point B be situated at the intersection of the wall and the straight element DB of characteristics of the second family drawn from point D and situated near the wall (Fig. 5.4.2). The magnitude and direction of velocity, entropy and coordinates of this point are known.
210 AERODYNAMICS
, The determination of velocity at point B is done directly with the help of the equation (5.4.4) for the characteristicsof the second family. If this equation is taken with the conditions along the element of characteristic DB and expressed in terms of finite difference,
where sin PDsin PD cos (BDPD)' sin2PD cos PD. cos (PDPO) ' (5.4.28)
rnD =
fD =
The coordinates XB, YB of point B are determined from the simultaneous solution of the equation of characteristics of the second family and the equation of wall contour:
A graphic solution of these equations is shown in Fig. 5.4.2, a. Angle PBcan be calculated from equation tan
PB=(dyldx)~
(5.4.31)
using the derived coordinates XB, y ~ Entropy SB(or stagnation pressure . pb,) at a point B is supposed to be known: it is equal to its value on the stream line coinciding with the contour of the network on which the entropy is supposed to be constant. After finding the increment AWD by equation (5.4.27) it is possible to calculate an angle COB=A ~ +WD and then Mach number MB can be deterD mined from Table 5.3.1. The scheme of graphic determination of velocity at a point B is shown in Fig. 5.4.2, b where the element D'B' of the characteristic of the second family in the hodograph corresponds to the element of characteristic of the same family in the physical plane. Third problem: This consists of the calculation of velocity at the intersection of characteristics with a shock wave and the determination of the deviation of shock at this point. As the characteristic happens to be the line of weak disturbances the above intersection physically refers to the intersection of a weak wave with shock. Let the adjacent expansion waves corresponding
METHOD OF CHARACTERISTICS
21 1
to the characteristics of the first family pass through points Sand H on shock
M N of given fo'rm y= f(x) (Fig. 5.4.3, a). As a result the intensity of shock
and hence the inclination of shock diminish. Since points S and H happen to be the sources of disturbances there arise reflected expansion waves and the characteristics of the second family can be drawn through these points. One such characteristic passing through point S intersects the neighboring conjugate characteristic at point F called the nodal point of characteristics. ,
a1
b)
"
Fig. 5.4.3. Scheme for caIculation of velocity at intersection of characteristics with compression shock:
aphysical plane; bhodograph plane; Icharacteristics of first family; 2characteristics of second family.
The properties of characteristics SF and FH passing through nodal point F and the relations for calculation of the shock wave are used to determine the variation in shock deviation and in the velocity behind the shock. As the length S H of the shock is small it can be assumed that this segment is a straight line. The angle of deviation of the shock on this segment and the corresponding gas parameters are approximately equal to their values at point of intersection H of the element FH of the characteristics of the first family with the shock. One of the unknown parameters is the angle of deviation of velocity vector /?H at this point, which can be represented as /?H= APF + PF, where A/?F = PH/?F and PF is the known angle of deviation of the velocity vector at point F. The second unknown, i.e. the Mach number MH at the same point, can be determined by the formula where (dM/d/?)~ the derivative which can be obtained from the theory of is shock for known conditions at point S. Quantity ADsH, which can be determined from the variation in angle /? along an element of shock, is equal to the difference PH Ps. This quantity can be taken as approximately equal to
21 2
AERODYNAMICS
the change in angle j? along element FH of the characteristic of the first fantily APFH=PB PF. The derivative d M / d p is calculated from the differentiation of (4.3.19'):
The derivative d(pz/pl)/dPcappearing on the right side can be obtained from the differentiation of (4.3.13):
The derivative dOc/dPc can be determined as foIlows: Let us differentiate (4.3.24) first:
1
sin 8c cos OC sin (8c  PC) cos (Oc  PC) 1 'sin (ec  PC) cos (Oc  PC) This relation can be slightly modified with the help of (4.3.24):
Equating the right sides of (5.4.34) and (5.4.35) and solving the equation so obtained for the derivative d6c/dPc, we find
Further, let us write the relation for variation in Mach number M during transition along the characteristic from point F to point H, i.e. for the quantity A M F = M H  M F . For this purpose the expression (5.4.32) can be used here in place of MH: (5.4.37) AMF =M s (dM/dP)sAPF MF.
+ +
It is possible to replace Mach number M with f~~nction obtained from the co relation (5.3.3 1): (5.4.38) AOIF os ( d ~ ~ ) / d P ) s  c o ~ , = APF
METHOD OF CHARACTERISTICS
213
where
Using the equation (5.4.11) for the characteristics of the first family to the element FH we find
where sin PF sin p~ sin2p~ cos PF. cos (PF w)' CF= cos (PF PF)
lp=
(5.4.42)
The distance between points F and H along the normal to the stream line An = (FH) sin p~ = sin p ~ . cos (PF  t p ~ )
XH  XF
wherep&, is found from the theory of the shock wave using Mach number
MH'
The entropy SF or stagnation pressure piF at a point F can be approximately taken on the basis of corresponding values at point S on the shock, i.e. SF%SS andpoFzpo,. Solving equations (5.4.38), (5.4.41) for APF, we get
I
)
Substituting the value of APF in (5.4.38), AWF can be found, the angle m~ =AmH+OF can be calculated and Mach number MH can be improved. After calculating the angle PH=APF +PF the new angle of inclination 8cH of the shock at a point H can be found from the value of this angle and the
givep Mach number M,. In this way its form on the segment SH is modiiied. If need be the calculations can be carried out taking values of parameters between points S and H in place of those at point S as a second approximation. and In particular the values of 0.5 ( o s + o ~ ) 0.5 (Ps+ PH) may be taken in place of os, PS respectively. The scheme of graphic solution is shown in Fig. 5.4.3, b. The point H' in the hodograph plane corresponding to the point H in the physical plane is determined as a result of the intersection of element F H of the characteris'' tic of the first family with the shock polar constructed for a given Mach number M I of an undisturbed flow. The vector O'H' determines the velocity AH at a point H.
5. Application of Method of Characteristics to Solution of the Problem of Contour Design of Nozzle of a Supersonic Aerodynamic Wind Tunnel
The method of characteristics helps us to solve one of the important problems of gas dynamics regarding the shape of the nozzle of an aerodynamic wind tunnel to obtain a twodimensional supersonic flow with a given speed. The nozzle providing this flow represents a type of tube with the front and rear walls plane, and the upper and lower walls in the form of a curvilinear specially profiled contour (Fig. 5.5.1).
Fig. 5.5.1. Nozzle of supersonic wind tunnel: 1upper wall; 2rear wall; 3exit section; 4lower 5front wall; 6critical section; 7reservoir. wall;
Along with the determination of the form of the curvilinear contour the calculation of gas parameters at the reservoir (parameters of complete stagnation) and at the critical section of area S* is included in calculations of the nozzle. The gas parameters at the nozzle exit such as Mach number M,, pressure p, area of exit cross section S and temperature TOare usually given quantities. The area of critical nozzle section is found from the equation of flow rate (3.6.44), which may be expressed in the form p,V,S =p*a*S*. From this
METHOD OF CHARACTERISTICS
2 15
From (3.6.46') i follows that the parameter. is determined by the given d Mach number M, at the nozzle exit. From this value of Mw and pressure p, at the nozzle exit the pressure required at reservoir po to achieve the given Mach number M, at exit can be found by the application of formula (3.6.29). In addition the angle 2y of a nonprofiled nozzle is given (Fig. 5.5.2). As shown by experimental research, this angle is usually selected in the range of 3035". Smaller values lead to an excessively long nozzle in which a thick boundary layer is formed on the wall, reducing the effective nozzle cross section. With higher values of the angle of the diffuser flow separation may take place, rendering the application of the nozzle meaningless in experimental research.
If the entrance section of the nozzle is smooth enough the motion in the region of the subcritical section can be regarded as an expanding radial flow from the source situated at point 0.Such a flow has the property that its direction coincides with the direction of radial lines originating in point 0. The variation in the magnitude of gas parameters along all these lines will have the same character. The length of the subsonic zone of a nozzle of unit width is determined by the quantity r*=S* [360/(2n x 2y x I)] and the distance to the exit section r~ = S[360/(2n x 2y x I)]. In the part of the nozzle restricted by the arcs of two circles with radii r* and r ~i.e. in the limit of the critical section, the gas flow will be supersonic. , The profiled nozzle has a straight wall BC in this part so that it provides gradual transition from radial flow to uniform rectilinear flow with a given velocity at the exit. For this purpose we draw a series of closely distributed lines from point 0 and determine velocity (Mach number M) on these lines at the points of intersection A1, A2, ..., A, of the characteristic of one of the families AA, (this will be considered as the characteristic of the second family), starting from point A on the arc of radius r~ (Fig. 5.5.3). Then point A1 lies at the intersection of the ray rl =OA1 with the element of characteristic
216
AERODYNAMICS
AA1 drawn at an angle of p,= sinl(l/M,), where M, is the given Mach number at the nozzle exit. Mach number M I at point A1 can be found with the help of the expression & =S*/SI. As S1=2nrl x 1(2y/360), S* =2nr* x 1(2y/360),we have & =r*/n. Then A1 and the corresponding Mach number MI can be found from (3.6.46). Similarly we can determine the coordinates of point of intersection A2 of neighboring ray r2 = OAz with an element of the characteristic AlA2 with an inclination of angle pl =sin'(1/M1) to the line OA1. The value of Mach number M2 and other quantities can be obtained as for the first case. As a result the characteristic of the second family m the form of segments of line AA1, AIA~,.An1 An, intersecting the network of straight lines BA, at point .. A,, can be constructed. As at other points A,2, A,I, An, etc. of intersection of characteristics with the straight lines from the source 0 , the corresponding Mach numbers and angles p=sinI (1/M) can be calculated at this point. The region of flow OAA, with the known velocity field bounded by the characteristic AAn and the straight walls of the nozzle is known as the triangle of information. The type of flow is maintained even if the nozzle contour behind point A, downstream is changed because the disturbances this produces cannot propagate upstream beyond the limit of Mach line AAa. Here a change of contour can be executed so that the radial flow on the Mach line is gradually transformed to a rectilinear flow of given Mach number M, at the nozzle exit.
Taking this condition into account, the characteristic of the first family emerging from point A in the form of a straight line can be constructed. Now if the velocity field is determined between the characteristics AA, and ADm the corresponding stream lines can be determined. The stream line that passes through point A, will coincide with the profiled contour of the nozzle. The condition that in the plane flow we are considering all the characteristics of the first family originating in points An2, An', etc. and the characteristic ADmwill be straight lines (see Fig. 5.5.3) is used to determine the velocity field. The inclination of each characteristic to the radial line is determined by the corresponding Mach angle ,UI =sin' (l/Ml), p2= sin' (1/M2),. .. etc. The
METHOD OF CHARACTERISTICS 2 17
velocity along the characteristic is determined by the value of velocity at the starting points Al, Az,... etc. Let us consider the stream line emerging from point A,. The initial part of this line coincides with the direction of velocity at point A, and represents by itself the line which is an extension of contour BAn until it intersects with the characteristic of the first family A,IA, at a point Dl. Behind point Dl the element of stream line coincides with the direction of velocityat point Dl, which is equal to the velocity at a point A,I. Drawing a straight line from point Dl parallel to ray OA,l up to the point of intersection Dz with characteristic A,2D2, we get the following part of the stream line. Behind point Dz the stream line on the segment DzD,2 (the point Dmz lies on the characteristic AzD,,z of the first family) will be parallel to the line OA,*. The construction of the remaining segments of stream line can be done the same way. The segment of stream line behind point Dm,lying on the characteristic AD,, is parallel to the nozzle axis. The contour of the nozzle, coinciding with the stream line A,D,+l and drawn in the form of a smooth curve, gives a uniform supersonic flow with a given Mach number M, at the nozzle exit. This method of profiling does not take care of the effect of the boundary Iayer on the form of the contour of a nozzle which distorts the character of the flow and changes the form of the stream line from the calculated one. The effect of the boundary Iayer can be accounted for approximately if the contour at the exit is inclined from the nozzle axis by an angle equal to 1020'. Using the equations of characteristics for a twodimensional flow, the construction of the contour of a circular nozzle can be carried out analogically to give an axisymmetric supersonic flow with a given velocity at the nozzle exit.
In this chapter problems connected with the formulation of aerodynamic theory for the calculation of flow around a wing profile in an incompressible fluid flow will be examined. The formation of a plane disturbed flow around an airfoil is the characteristic feature of this flowinteraction. Aerodynamic equations that are simpler than those for threedimensional flows are used in this investigation. Strictly speaking the flow around an airfoil that is supposed to be a plane flow is an idealized situation. In reality the flow around a real wing of finite span will be threedimensional. Therefore the aerodynamic characteristics of an airfoil cannot be used directly for a finite wing. However, these characteristics can be one of the basic parameters used in calculation of similar characteristics of real wings. Besides this, the solution of the problem of an airfoil has its own importance because the flow over certain parts of a wing is practically a plane flow. In the present chapter we will confine ourselves to the study of the interaction between an airfoil and plane incompressible fluid flow. At the same time the problem of a finite wing in the same type of flow will be examined. The results so obtained have their own significance and can be used (see Chapters 7 and 8) for aerodynamic study of highspeed motion.
1. Thin Airfoil in Incompressible Plow
Let us consider how to calculate the flow interaction of a steady incompressible fluid flow and a small, thin cambered airfoil at a small angle of attack (Fig. 6.1.1). The results to be obtained from this calculation for the aerodynamic characteristics of an airfoil can be directly used for cases of motion at low subsonic speeds (M, < 0.3 0.4) when the gas can be considered as an incompressible medium. These results can also be used as the initial data for aerodynamic calculations of airfoils of given form in a subsonic compressible flow.
21 8
As the airfoil is thin and the angle of attack is small the flow velocity near the airfoil will differ insignificantly from the undisturbed flow velocity. This kind of flow is said to be one with small disturbances. For the velocity of a small perturbation flow the following condition can be written: where u and v are the small perturbation components.
Upper contour
Y ,
= y,
(x)
According to this condition we have Now let us determine the pressure in a flow with small disturbances. From Bernoulli's equation in which C2 is assumed to be equal to (p,/p,) +(V2,/2) and p =const we get the change in pressure
P PW =  PW vw( a ~ l a x ) .
The corresponding coefficient of pressure will be
(..) 614
According to ( . . )the pressure difference on the lower side of the air614 foil will be pr pw =  (apl/ax) Vwp, and on the upper surface
220 AERODYNAMICS
where p,, pu are the velocity potentials on the lower and upper sides of the airfoil respectively. Consequently the lift force due to pressure acting on an element is dY=(prpu)dx=  V , p ,
a ~ l apu
(ax

ax)dx'
Let us consider the circulation of velocity along a contour having rectangular form with sides dx, dy surrounding an element of the profile. Then from Fig. 6.1.1 the circulation will be
Introducing the concept of intensity of circulation (vortex) determined by the derivative dr/dx= y (x), we get
As the slope dy/dx for a thin airfoil is small the product of this coefficient with the difference in the vertical components of velocity will be a small quantity of second order and hence In this way equation (6.1.6) can be written as
C~
From formulas (6.1.8) and (6.1.9) it can be seen that the coefficients of lift and moment due to pressure depend on the distribution of intensity of cir
221
culation over the profile. This shows that the flow past an airfoil can be replaced by a system of continuously distributed vortices. The verticalcomponent of velocity induced at a point with abscissa 5 (see Fig. 6.1.1) by the element of distributed vortex of circulation d r = y(x)dx may be given by the BiotSavart formula (2.7.12): dv =dr/[27t. (X
Then the velocity induced at this point by all the vortices will be
From the boundary condition over the airfoil surface we have v/(V, +arp/ ax)=dy/dx. Assuming that the airfoil is thin, Vm=arp/axwV, may be accepted and the derivative may be calculated from the given equation of mean line where y, =y, (x), yl= yr (x) are equations of the upper and lower contours of profile respectively. In this way the distribution of vortices y (x) is determined by the integral equation
Let us introduce a new independent variable 8 in place of x which is determined by the relation
We will find the solution of equation (6.1.12) in the form of Fourier's trigonometrical series
00
A. sin (n~~)],
(6.1.14)
in which the variable 00, according to (6.1.13), is connected with the coordinate x = 5 by the equation (6.1.15) Let us carry out the change of variables in (6.1.12). Differentiating (6.1.13), we get dx =(bj2)sin 8 d8. (6.1.16)
=(b/2) (1
cos 00).
+LC A.
n
n=l
CXJ
1
0
=i
0
1T
cos
0
cos 0 d8
=O; J
0
In the same work [l] it is shown that for the integer values* of m there is a formula
J
0
Consequently
{ n
2
AO+
Integrating both parts of equation (6.1.18) in the range of 0 to n, we get the relation for coefficient A0 in the Fourier series:
*These integrals are popularly known as Glauert's integral in the British and American literatureTranslator.
MuItiplying by cos 0 and then by cos 28, we find the expressions for the Fourier coefficients A1 and Az respectively. Thus in general
Let us consider the relation (6.1.8) for the coefficient of lift. Using the variable 0 therein and substituting (6.1.16), we get
0
Integration of this gives
sin 0 d0 +2
C I
cY=27~[Ao+(A1/2)1. (6.1.22) In this way the coefficient of lift depends on the first two coefficients of the series. Substituting expression (6.1.19) and formula (6.1.20) in (6.1.22) with n = l , we find
Similarly transferring the variable x to 0 in (6.1.9), for the coefficient of moment we get
* I
Using the expression (6.1.14) for y (B), the coefficient of moment is reduced to
Determining the values of A1 and A2 from (6.1.20) and substituting in (6.1.26), we get
Let us consider the coordinates X I , yl related to an airfoil, in which case the equation of mean line will be y, =yl(x1) and the angle of inclination of tangent will be given by the derivative dyl/dxl. This angle (Fig. 6.1.2) is =@ +a, where @xdy/dx. Finding the angle @ = P 1 a from this and inserting its value in (6.1.23) and (6.1.27), we get:
where
eo=
I
76
I
0
@I
(I  cos 8 ) dB;
I I !
4
Fig. 6.1.2. Mean line of profile.
From (6.1.28) it follows that the coefficient of lift is equal to zero at a=  8 0 . The angle a= eo is called the zero lift angle. The coefficients and ,UOare calculated from a given equation of mean line of profile under the I(, ;) condition that the variable in the expression for function / , is replaced in accordance with (6.1.13) by the relation 2, =(112) ( 1  cos 0).
225
Let us study an incompressible fluid flow over the simplest profile in the form of a thin plate, oriented perpendicularly to the direction of undisturbed velocity of flow. The solution of this problem may serve as an example of the application of the method of conformal transformation for investigating fluid motion. The results are used in real cases to determine the aerodynamic characteristics of finite wings (see Section 4). Let the plate be placed in the plane of a complex variable a =z + iy along real axis z (Fig. 6.2.1, b). The velocity vector of undisturbed flow Pwill coincide with the direction of imaginary axis y. In this case the flowinteraction of such a plate may be studied by means of conformal transformation from the flowinteraction of a circular cylinder in plane [= + iq (Fig. 6.2.1, a). If we know the function W=f(f:), representing the complex potential for flow about a circular cylinder, and the analytic function f:=F(o) of complex variable o=z+iy (the socalled conformed function), transforming the circular contour in place [ to the segment of a straight line placed in the flow in plane a in transverse direction, then the function W=f [F(o)] will represent the complex potential for flow around a flat plate.
Fig. 6.2.1. Conformal transformation of flow over circular cylinder (a) into flow over flat plate oriented vertically to velocity vector of undisturbed flow (b) over flat plate oriented along flow (c).
f: + (R2/0
(6.2.1)
gives this conformal transformation of a circle of radius R in plane f: to the segment of a line of length 2a=4R in plane o. The equation (6.2.1) can be rewritten in the following way:
As R2= +q2, after separating the real and imaginary components we get: z =25, y=O. Thus, the points on a horizontal line of length 2a = 4R in plane a correspond to the points on the circle of radius R in plane 5. Now we determine the complex potential for the flow around a circular cylinder. The method of conformal transformation, using the known function of complex potential for a flow around a flat plate oriented along the flow, is again used for this purpose. This function has the form W = q + i y = iV(z+iy)=  P a . (6.2.2) It can easily be shown that the conformal function (6.2.3) 0 =T  (R2/C) transforms the segment of straight line in plane a placed along the flow into the circle in plane [ (see Fig. 6.2.1, a and c). Practically speaking, it follows from (6.2.3) that since for the points situated on the line a = iy ( 2R ,<y 4 2R, z =O), we have
r2
Separating the real and imaginary parts of this equation, we get: q = ~ / 2 ,5= d R Z  (y2/4), whence C2+q2=R2. Thus the points lying on the circle in plane a correspond to the points lying on the vertical segment of line in plane T. Replacement of a in (6.2.2) by the value from (6.2.3) leads to the complex potential of flow around a circular cylinder of radius R in a rectilinear plane flow with velocity V:
To obtain the complex potential for the flow past a flat plate placed across the flow (see Fig. 6.2.1, b) the following value of obtained from the formula (6.2.1) is to be used in (6.2.4):
Then
As a =z for the points lying on the plate where a plex potential will be
where the plus sign refers to the upper surface and the minus sign to the lower surface. Calculating the derivative ay/az, the velocity past the plate and the pressure can be found. From the details obtained it follows that the pressure on the upper and lower surfaces of the plate is the same. Consequently in the case of an unseparated flow over a flat plate across the flow of an ideal (inviscid) fluid the plate has no drag force. This interesting aerodynamic effect will be studied in the example of the flowinteraction of a flat plate placed in a flow at some finite angle of attack in Section 3. The flow characterized by the potential function (6.2.7) is shown in Fig. 6.2.1, b. This flow can be represented as an irrotational flow around the plate which can be obtained by superposition of the flow from the dipole with potential W, =  iVc (6.2.8) to the undisturbed flow with potential
Wi = ~f: (R2/C), dp iV
where 1: is determined by the function (6.2.5).
3. Thin Plate at An Angle of Attack
(6.2.9)
Let us find the potential function for a disturbed flow of incompressible fluid past a thin plate oriented at an angle of attack a using the method of conformal transformation as in the previous problem (see Section 2). Let us place the plate in the plane of complex variable o = x + iy along real axis x. If it is assumed that the flow we are considering is irrotational its complex potential can be represented as the sum of potentials W 1 =V,u and Wz respectively for longitudinal and cross flows over the plate with an undisturbed velocity V=aV, (Fig. 6.3.1). The total complex potential will be
Taking into account (6.2.7) in which z is replaced by x, the total velocity potential past the flat plate will be
voc v ,
From (6.3.3) it follows that at leading edge (x= a) and trailing edge (x=a) the velocity Vx is infinite. Physically, this flow picture is not to be found. The limitation of velocity at one of the edges, for example at the trailing edge, can be found by superposing circulatory flow over the flow we are considering. The potential of rotating flow in plane is determined, in accordance with (2.9.22), by the expression W =  (ri/2n) In c. 3 Substituting here the value of [ from (6.2.5), we get (6.3.4)
Adding (6.3.1) and (6.3.5), we find the complex potential of rotationalrectilinear flow over an inclined flat plate: W= WI+ W Z +W ~ = V ~ Ciavm d o 2  4 ~ ~ T+
The circulation r can be determined from the ZhukovskiiChaplygin* hypothesis according to which the velocity at the trailing edge of a flat plate has a finite value. The value of this velocity can be obtained with the help of the theory of conformal transformation in the form of the derivative dF/da
*In the American literature this is known as the KuttaZhukovskii condition at the trailing edgeTranslator.
of complex potential W for a cylinder. This potential, in turn, helps us to find the complex velocity at the corresponding point on the cylinder in terms of the derivative dw/dl. By the rules of differentiation of a complex function this is equal to d@ .  do  = dW d l do dl' where and W are the complex potentials for the cylinder and flat plate respectively. By the ZhukovskiiChaplygin hypothesis the quantity dWldo is limited in its absolute value. As the derivative doldl, which can be determined from (6.2.1) by the formula
is equal to zero on the cylinder at a point l = R the derivative dF/dl=O at the corresponding point on the trailing edge of the plate. The complex potential for the cylinder will be
F=~ ~ + r n ~ + w ~ .
characterizes the flow in the direction of the axis that The potential corresponds to the flow over the flat plate in the same direction with a velocity V. It is obtained by replacing the quantity o from (6.2.1) in formula W ,= V,o:
WI = Voo I l (R2/l)1.
The complex potentials W2and F3 determined by the equations (6.2.4) are and (6.3.4) respectively. Replacing V by aV, in the first, we &d the total complex potential
For the point c= R corresponding to the point on the trailing edge of the plate the derivative dF/dl is equal to zero, i.e.
r=4naRV,,
230
AERODYNAMICS
o n substituting this result in (6.3.6)and differentiating with respect to a , the complex velocity will be given by
As 11 <a, we have x
dW
do from which it follows that the total velocity past the flat plate is
a  ~ ,  i V , = ~ , ( 1 ta,J&),  x
where the plus sign corresponds to the upper surface and the minus sign to the lower. The velocity at the trailing edge ( x = a = 2R) is equal to V and at , the leading edge (x=  a ) it is infinitely high. In actual cases the thickness of the leading edge is not equal to zero. Specifically, the edge of the airfoil, though not large, has a finite radius of curvature. For this reason the velocity at this edge has large but finite values. The general expressions for the main vector of forces due to hydrodynamic pressures acting on a steady cylindrical body of arbitrary shape in a steady incompressible fluid flow can be used to determine the forces acting on the plate. Let us introduce the concept of complex force F= X i by Y analogy with complex velocity, defining this force as the mirror image of the main vector 3 in the real axis. The vector F we are considering is determined by the formulas (1.3.2) and (1.3.3) in which the coefficient c, is taken to be zero. Then
23 1
where n is the direction of the external normal to the contour C of the body in the flow; 0 i s the angle between the element of contour ds and the x axis (Fig. 6.3.2).
Since da=dx+idy=ds (cos e+isin O)=eie ds; d a = dx  idy = ds (cos 0  i sin 0) =e'@ds, and the pressure is determined by the Bernoulli equation:
(6.3.14)
it folIows that
Taking into account that dG=e2ieda on the basis of (6.3.14) and keeping in view that the complex velocity at the given point of contour by the condition of unseparated flow

V=V, ivy = V c o ~ 0  i V s i n 0 = V e  ~ ~ ,
(6.3.16)
we find
The integration in (6.3.18) can be carried out over other contours surrounding the given contour C. As an example of such a contour the circular contour K may be selected, so that the complex velocity in plane o can be expressed in the form
where Vm=v,, ivy, and the coefficient A can be determined from (6.3.8'). The square of the complex velocity will be
Here the first and third terms are zero. The integral $ do/o is calculated taking into account the formula a =x + iy = reiq (see Fig. 6.3.2) and is equal to
K
Coming back to the conjugate value of this force, we have where V,=Vxm+iVyW. According to the expression for the force acting on a contour it is equal to the product of the density, velocity of undisturbed flow and circulation of velocity:
This expression is known as the Zhukovskii formula. From (6.3.21) it follows that the acting force represents the vector that is perpendicular to the velocity vector of an undisturbed flow. Hence it is a lift force. Its direction coincides with the direction of the vector obtained by
turning the velodty vector 7, through an angle of x/2 against the direction of the circulation. This force is called the Zhukovskii force. Let us apply formula (6.3.22) for calculation of the lift force on a flat plate. Keeping in mind the value of circulation from (6.3.9'), we get The normal and axial components N and R of the Zhukovskii force are respectively determined as follows (Fig. 6.3.3):
N = Y cos a % Y= 2naapV2;
=  Y sin a w
(6.3.24) (6.3.25)
 Ya = 2na2apVi.
The presence of axial force R acting along the flat plate against the flow seems to offer a paradox because all the elemental forces due to pressure are directed along the normal to a surface. This component of the Zhukovskii force is known as the suction drag. The physical nature of its existence is explained like this: Let us assume that the leading edge is slightly rounded. Then the velocities in the neigh OL borhood of the leading edge would be high but not infinitely high as happens in the case of a flat edge' Fig. 6.3.3. System of forces acting on Bat According to Bernoulli's equation the plate (Rsuction force). difference between the pressures at the leading edge and at infinity will be negative. Separation also results in the suction drag. The value of the suction drag is given by the expression (6.3.25). The expression (6.3.25) for suction drag refers to the case of flow past a flat plate where the velocity near its leading edge is determined by the formula (6.3.12). The axial component of perturbation velocity about a given plate, determined by this formula, will be
The expression for suction drag can be generalized for the case of an arbitrary given velocity near the 1eadi.ngedge. For this purpose we write (6.3.25) in the form where
c2=2aa2V2.
Quantity C2 can be represented in the form of the expression
where XL.E is the abscissa of leading edge (XL.E a); x is the varying abscis= sa of the points on the flat plate; V i is the axial component of perturbation velocity on the upper side of the wing. It can be shown that formula (6.3.28') is valid in practice. To do this we introduce expression (6.3.26) in (6.3.28'):
Substituting this quantity in (6.3.27), we get the formula (6.3.25) obtained earlier. The results obtained for an incompressible irrotational uniform flow over an airfoil in the form of a thin plate are useful in studying the aerodynamic characteristics of airfoils of actual shapes and also for finite .wings in incompressible fluid flows as well as in a compressible gaseous medium. 4. Finite Wing in Incompressible Fluid So far we have been examining the flowinteraction of an airfoil and an incompressible fluid flow. It can be assumed that this profile corresponds to an elemental part of the lifting surface of a wing of infinite span in a rectilinear uniform flow. Hence the theory of this flowinteraction also happens to be the foundation of the aerodynamics of wings of infinite span. According to formula (6.3.22) the lift force on the part of a wing of unit span Y(')=p,V,T (Fig. 6.4.1, a). Consequently there is a rotational flow with circulation r around b) the profile. If the circulation is in. a clockwise direction . the velocities will be higher ' + I= K. on the upper side of 'the profile (the circulatory flow Vortex of ~ntensity X is superposed on the oncoming flow in the same direction) and lower on the under side of the profile (circulatory flow is in the opposite direction to the undisturbed flow). Therefore, by Bernoulli's equation the pressure will be Bound vortex Free vortices less on the upper side than On the lower side and the Fig. 6.4.1. System of equivalent vortices for wing of rectangular plan. lift force will be directed
. \
. . .'.
 L 
upward as shown in Fig. 6.4.1, b. Since the circulation, according to (2.7.8) and (2.7.8'), is equal to the intensity of vortex K the element of wing can be replaced by the equivalent vortex with a given intensity of vortex along the span. Zhukovskii called this vortex the bound vortex. In a hydrodynamic sense a wing of infinite span is equivalent to a bound vortex. Let us now consider an approximate scheme of flowinteraction of a finite wing with a rectangular wing plan. The bound vortex, as established by S.A. Chaplygin, turns near the wingtips in the form of a pair of vortices and moves away behind the wing in the direction of the undisturbed flow. The distance e (see Fig. 6.4.1, c) from the point where the vortex turns to the wingtip depends on the geometry of the wing. Hence the hydrodynamic effect of a finite wing can be obtained by replacing it with a bound vortex and a pair of free vortices resembling the form of the letter l .This scheme of wing is 7 known as Chaplygin's 17shaped scheme. The equivalent vortex system of a finite wing induces additional velocities in flow and results in the downwash that is the characteristic of flow past a finite wing. The following theorems of Helmohltz form the basis for calculation of the induced velocities and downwash angle due to free vortices: 1. The intensity of a vortex does not change in magnitude and consequently a vortex cannot suddenly cut off or terminate in a fluid; 2. The intensity of a vortex is independent of time; 3. A vortex in an ideal fluid is not subject to extinction. In the system of a rectangular wing that we are considering the circulation along the span is taken to be constant on the assumption that the lift force of each elemental segment of the wing is equal. In reality the lift force of this rectangular wing plan varies along its span. This variation is small in the middle part of the wing but quite significant near the wingtips. For a wing of arbitrary plane the variation in circulation has an absolutely clear form which is based on the unequal sizes of segments and hence is characterized by different values of lifting force. The vortex system of flow over a wing not of rectangular plan can be obtained if the wing is replaced not by a single 17shaped vortex* but by a system of 17shaped vortices forming avortex sheet (Fig. 6.4.2). The circulation will be constant along each vortex but will change from one vortex to another. For the middle cross section of the wing (root section) the lift force is maximum and therefore the intensity of vortex and circulation will be maximum at that section. Experiments show that this vortex sheet behind the wing is unsteady and at a comparatively small distance from the wing it twists around in two parallel vortex threads (see Fig. 6.4.2).
*In the English and American literature it is known as a horseshoe vortex systemTranslator.
236 AERODYNAMICS Let us study the relations required for calculation of induced velocities and'downwash angle. First we have to determine the distance 10 between the free (curling) vortices. Here we will assume that the vortex system for a wing of span I can be replaced by one horseshoe vortex with constant circulation T O . It is assumed that the fixed vortex, known as the lifting part of a horseshoe vortex, passes through the aerodynamic center of a wing with a coordinate x,.,.
The distance 10 between the free (curling) vortices is determined in the following way: The area under the curve T ( z )is equal to TavIand the average circulation rav along the span is determined from the known lift force Y on the wing with the help of Zhukovskii's formula Y=p,V,Tavl. In accordance with this formula, taking into account that Y = c,q,bavl, average circulation will be Therefore for circulation it may be written:
where c,o is the coefficient of lift for root section; b, is the root chord. Equations (6.4.1) and (6.4.2) establishing the relation between circulation of velocity and the corresponding values of the coefficient of lift and wing chord are known as theequations of correlation. From these equations we have
The elemental value of the lift force, by equation (6.3.23), will be dY=p,Voordz and the lift acting on the wing wit1 be
237
Y=2p,Vw
The concepts of vortex systems rav TO our assumption correspond and on to one and the same lift force Y and therefore
f
0
rdz.
(6.4.4)
where wl is the velocity at a point A induced by the bound vortex. According to (2.7.11)
wl = (  r o / 4 w ~ )cos PI cos ( n  P 1 ) ] = (  r o / 2 n L ) cos 81. [
Therefore
e l = T O cos P1/2nV,L=r0 cos2Pl/wVmsin pilo. (6.4.8)
The induced velocity w: at the same point A due to one of the free vortices is calculated by formula (2.7.11). It is necessary to make the following substitutions: r =To, h= 10/2, a , =0, a2= w  8 2 . As a result
W: =( ~o/[4w(lo/2)]}(1 .P2)= (  r0/27dO)(1 cos
+sin B,).
(6.4.9)
The velocity created by a pair of free vortices will be twice wi: w2= 2wi= ( r o / ~ z o )+sin PI). (l The corresponding downwash angle is
82 =  w2/Vm= (ro/wVWlo)(1 +sin
(6.4.10) (6.4.11)
P 1).
The total value of the downwash angle due to the system of horseshoe vortices is
r o ( l +sin 81) & = & 1 f & 2 =  (c0s2P1+1+sin81)= nVW1o sin /J'I wV,Zo sin '
(6.4.12)
238
AERODYNAMICS
we get
8 = (cyob,/2n~o)[l dl
+ +(IO/~L)~].
(6.4.13)
According to (6.4.6)
k =1/10, 1= l/bav ,
and taking into account (6.4.14), equation (6.4.13) can be transformed into the form
According to formula (6.4.13') the downwash angles behind the wing in a subsonic flow are proportional to the coefficient of lift and are inversely proportional to the aspect ratio A. Let us introduce the new quantity
then
Research has shown that the quantity z is less than unity and therefore the downwash angle can be regarded as a function basically determined by the coefficient of lift and the aspect ratio of the wing. Now let us see what types of changes are brought about by downwash in the flowinteraction of a wing at an angle of attack a (also known as the goemetrical angle of attack) in a flow with velocity 7,. The appearance of downwash at angle e behind the wing results in the flowinteraction being characterized by values of velocity and angle of attack a, differing from the values of Vrn a that deterand mine the flow around a wing of infinite span. The effective angle of attack a, of a finite wing will be less than the geometrical angle of attack by the value of downwash angle e (Fig. 6.4.3), i.e.
The presence of downwash leads to a change in the forceinteraction of the medium and the body in the flow. If the downwash were absent the vector of an aerodynamic force, according to formula (6.3.21), would have been normal to the direction of undisturbed velocity In the presence of
vw.
downwash velocity the resultant aerodynamic force vector will be oriented along the normal to the direction of reaI velocity p (see Fig. 6.4.3). As a m result of inclination of the resultant force by angle e there appears component Xi in the direction of undisturbed flow. This complementary force Xi, appearing as a result of downwash induced by the vortices, is known as the induced drag. From the physical point of view the presence of induced drag is associated with loss of part of the kinetic Fig. 6.4.3. Downwash near wing and energy of a moving wing, spent in occurrence of induced drag. the formation of vortices shed from its trailing edge. The value of this drag is determined from Fig. 6.4.3 by the expression
"_
Here lift force Y is found in the same way as for the wing of infinite span on i the assumption that the downwash is small. If the force X is divided by the quantity (p,Vi/2)S,v we get the coefficient of induced drag If we assume the values of k e 1, (1/2L)2/k2e1in expression (6.4.13') for E and introduce notation 1 =A,, then, after substitution in (6.4.18), we get
1. Subsonic Compressible Flow Over Thin Airfoil 1.1 Linearization of equation of velocity potential The study of the flowinteraction between an airfoil and a subsonic flow is related to the solution of the equation of velocity potential for a plane twodimensional flow that is obtained from (5.1.8) at & =0. This has the form
This partial differential equation of second order happens to be the nonlinear equation with respect to the unknown function 9. It describes the flow past fairly thick profiles creating large disturbances in the gas such that velocity of flow V and speed of sound a differ considerably from the corresponding parameters of an undisturbed flow. If an airfoil is thin and the perturbations it creates are small, the equation (7.1 . l ) can be simplified to the linear differential equation with constant coefficients and secondorder partial derivatives. This simplification is known as linearization and the equation so obtained and the small perturbation flow described by it are known as the linearized equation and linearized flow respectively. The velocity conditions (6.1.1) and the actual equality (6.1.2) are fulfilled for a linearized flow. Within an accuracy of the order of perturbation velocities u and v (6.1.1) the equation for the speed of sound can be formulated from (3.6.20), which takes the form
a2=aL [(k 1)/2]( V L  V2).
(7.1.2)
Using the expression (6.1.2) here in place of V 2 ,we get Substituting the quantity (7.1.2') in (7.1.1) and recalling that Vx=Voo+u, Vy=v, v ; = v : + 2 v o o U, v ; z v 2 ,
240
and further a s s u ~ i n g the total velocity potential of a linearized flow can that be expressed by y, = y,, ip', where p, is the velocity potential of an undisturbed flow and the perturbation velocity will'be y,'<y,, according to the condition (6.1. l), we get
The secondorder partial derivatives of the potential y,' with respect to the coordinates x, y are small quantities of first order:
Taking this in (7.1.3), the group of small terms of second and third order can be determined and, neglecting them, we get the linearized equation in the following form:
where M, = V,/aw. Let us study the expression for pressure in a linearized flow. For this purpose we use the formula (3.6.26), which will be rewritten as
k/(k1)
we find
Expanding the right side of this expression in series and maintaining the second term therein, we get
From this the pressure differencewill bep p, =  p,V,u and the coefficient of pressure will be ?= 2u/V,, i.e. we get the same relations (6.1.3), (6.1.5) as in the case of an incompressible fluid flow. However, in applying these relations in the case of high subsonic speeds it is necessary to take into account that the perturbation velocity u = ayr/ax should be determined taking into account compressibility.
1.2 Relationship between compressible gas flow and incompressible fluid flow over thin airfoil The flowinteraction of a thin airfoil in a compressible subsonic flow at a small angle of attack is studied with the help of equation (7.1.4') in which M, < 1. Let us now perform the replacement of variables in accordance with the relations
xo=x, y o = y d l  ~ %, v;=P'Y(V,O/V,),
(7.1.5)
where y is some arbitrary parameter, VO is the velocity of a fictitious flow , , (fictitious velocity) which differs in general from the velocity V of a given flow. The substitution of (7.1.5) in equation (7.1.4') gives the Laplace equation for determination of velocity potential p; for an incompressible flow in plane xo, yo:
In this way the problem of flowinteraction between a given profile and a compressible fluid flow can be solved using the results of the solution of the problem of an incompressible fluid flow with a fictitious velocity VO over , some transformed airfoil. Let us find out the relation between the corresponding parameters of flow past an airfoil and their geometrical characteristics. The relation between perturbation velocities uo and u in incompressible and compressible flows respectively is established, in accordance with (7.1.5), in the following way:
Substituting (6.1.5) in this equation, the coefficients of pressure in incompressible and compressible flows 5,= 2uo/V,o and p =  2u/V, respectively can be found. Consequently from (7.1.7) we get Pin= YP. From the formulas
Cy
(7.1.8)
in = $Pin
TO, m~ i n =
,#En
QO
TO
along with the expression (7.1.8) we find the relation between the corresponding coefficients of lift and moment:
Let us establish the relation between the forms of the airfoils and the angles of attack. For this purpose we first determine the relation between the vertical components of velocity. According to (7.1.5) we have:
For an incompressible fluid the condition of unseparated flow (3.3.17'), by o which F= F =yo fo (xo) and yo= fo (xo) as the equation of airfoil contour, results in
or, recalling that uo< Vwo, it will be Similarly for a compressible gas flow it will be Consequently
Under the condition that at y=O the quantity yo0; integration of this relation gives the equation connecting the vertical coordinates of fictitious and given profiles:
At the same time it follows from (7.1.5) that the horizont.al coordinates of these profiles do not change. Considering this, the angles of attack in the above two cases can be expressed in the form
a i n =yo/(b
where'b is the distance to the trailing edge of the profile, x is the horizontal coordinate of a point (Fig. 7.1.1). Thus according to (7.1.11) we have
a i n =a(y/
41 M. ) :
(7.1.12)
'
=F
~n 7
y Z Y o JIMZ,,
Cy = Cy
a=ain 41 M L ; ~ 
(7.1.13)
In this way, if the pressure coefficients at the corresponding points of the profiles in the compressible and incompressible flows are identical the profile in the compressible flow must be thinner than that in the incompressible flow by 41 ML times. The angle of attack will be reduced in the same proportion. 
Fig. 7.1.1. Profiles in smallperturbation incompressible (a) and compressible (6) flows.
From the results obtained it can be seen that for the two profiles under the same angle of attack the pressure coefficients at the corresponding points of the profile and their total lift coefficient and moment coefficient in the compressible flow will be l / d l  ~ 2 ,times greater than those in the incompressible flow. From this it can be concluded that the compressibility effect leads to a rise in pressure and lift. The factor 1 / d 1  ~ 2 is called the PrandtlGlauert correction factor for compressibility effect.
2. Academician S.A. Khristianovich's Method
2.1 Description of method If the airfoil or other body under flowinteraction induces finite disturbances in a flow the linearized equations are not applicable. In studying this kind of flow it is necessary to use the nonlinear equations of gas dynamics. Many problems of supersonic flows of gas can be solved by the application of the method of characteristics. The number of problems in supersonic aero
245
dynamics that can be solved is further increased by the numerical methods of integrating equations of motion. The analysis of a subsonic gas flow is quite difticult. Mathematically this can be explained by the different types of equations: hyperbolic for supersonic flows and elliptical for subsonic flows. The application of imaginary characteristics for elliptical equations does not yield any particular simplification. The greater difficulty in the investigation of subsonic flows can be physically explained by the fact that the disturbances in such flows propagate all over in the region of motion. On the other hand, in the case of supersonic flows the disturbances are confined by the surface of a cone with its apex at the point of disturbance and they propagate only downstream. Here the investigation of smallperturbation subsonic flows, due to the linearized character of the equations, is somewhat simpler than for subsonic flows with large perturbations during the Aowinteractions of thick profiles. Academician S.A. Chaplygin's book on gas jets is devoted to the investigation of the method of studying such flows. In it the equations forming the mathematical basis of the theory of potential subsonic flows are set out. These equations are known as Chaplygin's equations in gas dynamics. Their specialty is that they determine the motion of a gas not in the x, y plane but in the plane of special coordinates z and /3 (z = V 2is the square of total velocity at a given point in flows; p is the polar angle which can be determined from V = V x cos p). Unlike the usual equations they are linear because the coefficients in these equations represent by themselves the functions of independent variables z, P. These equations were solved by S.A. Chaplygin for a series of cases of gas motion at high subsonic speeds. Chaplygin's equations turned out to be the basis for many other methods in the field of highspeed aerodynamics. Using these equations, Academician Khristianovich worked out a method that helps us to characterize the effect of compressibility for a subsonic flow past airfoils of arbitrary shapes. The theoretical concepts of this method are to be found in detail in work [25]. Let us study the basic concept of this method and its application to the solution
Fig. 7.2.1. Calculation of pressure over profile in compressible flow: areal flow; 6fictitious flow; Igiven profile; 2fictitious profile.
of'various problems of flowinteractionsof profiles in a compressible subsonic flow. Examining a compressible flow past an airfoil of given form, S.A. Khristianovich showed that the equations of this flowinteraction can be assimilated to the equations of an incompressible fluid flow past an airfoil of some different form (Fig. 7.2.1). So, with the help of S.A. Khristianovich's method a relatively simple problem of a fictitious incompressible flow past some fictitious profile is solved to start with and then the parameters obtained are worked out for the conditions of the compressible flow past a given profile. The above analysis is based on the use of the functional relation between the real relative velocity L= V/a* in the compressible flow and the value of the fictitious nondimensional velocity A = Vola* at the corresponding points on the actual and fictitiousprofiles respectively (Table 7.2.1). Here the method we are considering makes it possible to transform the given profile into the fictitious profile.
........................................
A
A
0 0 0.05 0.0500 0.450 0.4307 0.10 0.0998 0.15 0.1493 0.20 0.1983
TABLE 7.2.1
0.25 0.2467
          A
A
0.500 0.4734
6.550 0.5144
0.600 0.5535
0.30 0.2943
0.625 0.5722
0.35 0.3410
  0
0.650 0.5904
0.7000 0.6251
7: : : :
         
0.825 0.6988
0.850 0.7110
........................................
As shown by S.A. Khristianovich, the difference in the forms of the actual and fictitious profiles can be neglected in the case of thin airfoils. S.A. Khristianovich's method allows us to recalculate the flow parameters over the airfoil (pressure, velocity) fairly easily for any Mach number M,>O, i.e. accounting for the compressibility, if the distribution of these parameters around the same airfoil in a low subsonic flow without compressibility effect (M,xO) is known. In addition, this method helps us to extend the calculation of parameters for one Mach number M,, > 0 to the other Mach number M,z >O (M,z#M,I). I n performing calculations by S.A. Khristianovich's method it is necessary to bear in mind that it is valid under the condition of subsonic flow velocity all over the profile. This condition is observed when the Mach number of the undisturbed flow is less than its critical value Mwcr. Therefore before embarking on the calculation it is necessary to find this critical value and to
0.925 0.7413
247
establish that &<&,,. The critical Mach number M,,, termined by S.A. Khristianovich's method.
2.2 Conversion of pressure coefficient from incompressible fluid flow to Mach number M, >0 Let the distribution of the pressure coefficient over an airfoil in an incom pressible flow be known, i.e. the nature of functionposp~ be known (Fig. (x) 7.2.2). The transformation of this function for the new Mach number a >M, > 0 will be carried out in the following order: ,
Fig. 72.2. Type of pressure distribution on one side of profile at different values of Mw.
The fictitious speed A , of an incompressible flow corresponding to the value of I, can be found from Table 7.2.1. For a selected value of from Bernoulli's equation
6,
Po=
Knowing A, the actual local velocity I for a compressible flow can be obtained from Table 7.2.1 and the pressure p can be calculated by the formula (3.6.26) in which it is assumed that V2/V2,,= [(k I)/(k + I)] I*. The pressure coefficient is then determined by the formula p= 2[(p/pm) l]/lzML. The
nature of curve p=p (x) recalculated for a given Mach number M, is shown in Fig. 7.2.2.
2.3 Conversion of pressure coefficient from Mach number >0 to Mach number M,z >0 (Mm1#&2)  Let us assume that the distribution of pressure coefficient p~=pl ( x ) is known in the compressible fluid flow at some Mach number M,I (Mwc,> Mwl >0). To calculate the corresponding distribution of the pressure coefficient at the new Mach number Mw2> O it is necessary first to find the relative velocities 11 and ,?,2 by using formula (7.2.1) for the Mach numbers M,I and MW2 respectively. Then the relative velocities Awl and A,2 of a fictitious incompressible flow can be determined from Table 7.2.1. By inserting some value of coefficient of pressure the absolute pressure
FI
can be found from the formula & = 2(pl p,l j/(kMmlpWl)and the local relative velocity of the actual flow can be obtained:
a*= k+
1 
( )
(k I)/IC
]}li2
For the value of ,?I Table 7.2.1 gives the local velocity in the fictitious incompressible flow. The corresponding coefficient of pressure can be calculated by Bernoulli's equation (7.2.2) Calculation of coefficient of pressure from this value of 50 Mach for number MW2 carried out in the same way as the conversion of the pressure is coefficient of the incompressible flow to the Mach number M, >O already discussed.
2.4 Determination of critical Mach number M According to S.A. Khristianovich's hypothesis, in an incompressible flow past an airfoil the local sonic speed corresponding to the critical Mach number of an undisturbed flow M,,, arises at the point where the highest suction occurs. Khristianovich established the relation between the minimum pressure coefficient Pomin for the highest suction and the critical Mach number M,,,. So it is clear that the value of the highest suction pornin should be determined first by some means, for example with the help of a lowspeed wind tunnel, before attempting to find the critical Mach number. If the pressure distribution, taking into account compressibility, for M,,,>M,>O is obtained from the wind tunnel test the quantity &,,in can be ,determined by the
reverse conversion of highest suction ;omin to the Mach number M,=O. Let us assume that the value of highest suction pod, is known. As the relative velocity A= 1 at the point of appearance of sonic speed, the value of the local velocity of the fictitious incompressible flow A=0.7577 from Table 7.2.1 for the value of I=1. Now from Bernoulli's equation
Pornin = 1  (A/4/1,)2
the relative speed of the fictitious undisturbed flow will be and Table 7.2.1 and the value of A give the critical velocity A, in the com, , pressible flow. The corresponding critical Mach number will be
The graph of critical Mach number versus the results of the above calculation is shown in Fig. 7.2.3.
3. Wing Profile in Flow Having Supercritical Velocity (M, > MWo)
0.8.
A subsonic flow past an airfoil may be . characterized by the following two cases. In 0 5 the first case the local flow velocity on the 0.40.5 1.0 1.5 Gmin surface does not go beyond sonic velozity anywhere on the airfoil. This is a case of pure ~ i 7.Ze3. . ~ curve determinafor subsonic flow interaction. It is known that at tion of critical Mach number M this velocity shock waves cannot arise at any by S.A. Khristianovich. point on the airfoil. The lift and drag forces will be determined with the compressibility effect. In general they depend on the forces due to normal pressure and friction. This kind of drag consisting of drag due to normal pressure and frictional drag is known as profile drag. In the second case, known as the supercritical flow interaction, the Mach number M of an undisturbed flow is higher than critical, i.e. M, >M,,,. The local velocity at some points around the profile becomes higher than sonic velocity and a zone of local supersonic speeds comes into the picture. Now the flowinteraction of the profile is characterized by the fact that the local velocity is less than sonic velocity ahead of the leading edge as well as behind the airfoil. Therefore during transition from supersonic to subsonic speeds local shocks arise near the wing (Fig. 7.3.1). These shocks, resembling in form the letter I ( I shocks), consist of two shocks each: the front curvilinear (oblique) shock DE and the rear almost linear shock CB (see Fig. 7.3.1, b).
' Local shocks, which may appear on the upper as well as the lower surface of the wing profile, significantly alter the pressure distribution and set up additional drag known as wave drag Xw.Thus the total drag of profile X will and i.e. consist of profile drag Xp, wave drag Xw,
x= XPI+x w
or
Cx = Cxpr
+Cxw,
where c, is the coefficient of resultant drag of an airfoil; c and c,, are the , coefficientsof profile and wave drag respectively.
I'
nl
Fig. 7.3.1. Profile in subsonic flow with supercritical speed: ascheme of calculation of flowinteraction in presence of local normal shock; bpressure distribution over profile with formation of local 1 shock.
Let us examine the simple method of calculation of wave drag suggested by Prof. G.F. Burago (see the book [I]). Assume that a local supersonic zone is formed on the upper side of the profile which takes the approximate form of the nearly straight shock BC (see Fig. 7.3.1, a). Let us separate out a flowjet passing through this shock. The gas parameters in this jet will be V 1 , ply plyMI just before the shock and V2, pz, M2 behind the shock. Take two p2, control surfaces II, 1111 to the left and right side at a fair distance from the profile. Denote the gas parameters on the left plane by V1,, pl,, pl, and those on the right plane by V2m, ~ 2 ~ ~ 2 9 ~ . Applying the theorem according to which the rate of change of mass momentum of a gas flowing through the control surfaces is equal to the resultant force acting on the left and right surfaces and the wave drag force Xw acting on the profile, we get
where dml,, dmz, are the mass flow rates per second of gas along the jet:
These are equal from the condition of constant flow rate, i.e. (7.3.2) plm Vim d ~ l m = ~ ~ w V2x dy2m. As postulated by G.F. Burago, we assume that equality of velocity is reestablished at a considerable distance behind the wing, i.e. V2,+ Vim. Accordingly, using (7.3.2) we get
Consequently
It can be assumed that p2, =pi, for jets not intersecting with the shock, i.e. the pressure at a considerable distance behind the profile settles down to the value of pressure in the undisturbed flow. For jets passing across the shock we have p b <pi,. In reality, since
P ~ ~ I P ~ ~ = P ~ I P ~ =< 1. VO (7.3.5) Taking into account (7.3.9, equation (7.3.3) can be transformed with the help of the equation of mass flow rate
plrnvlcod~lrn=p~V~d~Js, so that
252
AERODYNAMICS
where s is the length of the shock. The function vo (4.3.20) may be expressed in the form of a series with n=Ml 1:
'In the expression (7.3.7) the value of vo (O)= 1 because at M I = 1 (n=O) the shock changes to a wave of infinitesimally small intensity and the pressure
pb=po.
Keeping in view that for thin airfoils at small angles of attack the difference (MI 1) is small and in the series expansion (7.3.7) it can be truncated to the fourth term, we have
As shown by experiment, up to the given approximation it can be assumed that the quantity (MI I) for a given airfoil is proportional to the difference (M,MmCr). Denoting the corresponding coefficient of proportionality by A1 and including it in the general coefficient A, which can be determined in the form
we get the following expression for the coefficient of wave drag using (7.3.6') and (7.3.7'):
Cxw  A
 (Mm  Mmcr)3.
(7.3.9)
The coefficient A, in the general case, depends on the form of the profile, angle of attack and Mach number M,. However, aerodynamic experiments in wind tunnels on modern airfoils at low angles of attack showed that coefficient A z l l . In such cases satisfactory results are obtained from formula (7.3.9) if the difference M, M,,, does not exceed 0.15.
253
From formula (7.3.9) it follows that the coefficient of wave drag increases with the rise in M. This happens because the shocks appearing at the profile become more intensive and elongated Cx with the rise in flight speed. To reduce c x , at a given M, it is neces U.5 sary to raise the critical Mach number 0.4 MW,, which is basically achieved by reducing the thickness of the profile. 0.2 A similar effect can be achieved by 0. r reducing the angle of attack. Figure 7.3.2 shows the empirical 0 0.2 O,j 0.4 0.6 o,7 curve characterizing variation in coFig. 7.3.2. Profile drag in transonic zone efficient of drag c =c,,, +c,, in the , of flowinteraction. transonic regime of flowinteraction. For values of M~0.450.5 there is only profile drag. The higher values of M, correspond to supercritical velocities of flow (M,>0.50.55) where wave drag comes into the picture due to the formation of local shocks.
@,
4. Flat Plate in Supersonic Flow of Gas Having Constant Heat Coefficients Let us study the simplest wing profiIe in the form of a flat plate placed in a supersonic flow at an angle of attack a. The scheme of flowinteraction of this plate is shown in Fig. 7.4.1. Near the leading edge the supersonic flow is split into two parts: upper (above flat plate) and lower (below it)parts which do not affect each other. Therefore the supersonic flowinteraction on each side of the plate can be studied independently. Let us examine the upper side of the plate. The flow over this side represents in itself a plane supersonic flow over a surface which makes an angle of more than 180" with the direction of the undisturbed flow. This kind of flow, shown in Fig. 7.4.2 with the plane O C corresponding to the upper surface of the plate, was first investigated by Prandtl and Meyer and is called the PrandtlMeyer flow. According to the system of flow in Fig. 7.4.2 the flow parallel to the plane OB gradually turns, undergoes expansion and takes a new direction parallel to the plane OC. The angle of inclination of this plane Doc to the vector F, corresponds to the angle of attack a of the flat plate in Fig. 7.4.1. The disturbed part of the flow with expansion is limited from the left side by the Mach line O D which is inclined undisturbed , at an angle p,= sin'(l/M,), to the velocity vector ofnumber in theflow V where M, is the Mach undisturbed flow. The process of expansion ends on the Mach line O E inclined to the velocity vector of disturbed flow VOc by an angle pot= sin' (l/Moc), which can be determined from the Mach number of the disturbed flow over plane OC. The change in the direction of flow between the Mach
lines OD and OE can be expressed as the successive summation of inclinations of stream lines by small angles A p . To each such inclination indicating the appearance of additional disturbance there corresponds a straight Mach line passing through point 0.
In this way the region of turning will be filled up by an infinite number of Mach lines forming a "fan" of lines of disturbance which characterizes the centralized expansion wave. This centralized wave, known as PrandtlMeyer's expansion fan, is determined by the straight Mach lines. The flow parameters are constant along each of these lines and therefore each line corresponds to a series of plain expansion wa7res. The problem of the motion of a gas about an obtuse angle, which is related to the formation of a centralized expansion wave, can be solved by the characteristics method. The point of intersection F of the stream line in a steady uniform flow (angle of inclination of the stream line at this point /3 =0) with the characteristics OD in a fixed plane corresponds to point F' on the epicycloidal characteristic in the hodograph plane of the same family. To be more precise, each of these characteristics can be referred, for example, to the characteristics of the first family. The equation of characteristics of this family in the hodograph plane will be P =o t PI. Since p = 0 by the above where the angle o is found from (5.3.31) , condition we have P I = o,(M,), for a given Mach number M,. Hence the equation of characteristics will be P =o  ow, which from For a given small inclination of stream line P = A p the corresponding angle can be taken as w = AD o and the Mach number M can be found out , on the neighboring Mach line inclined to the new direction of the stream line
at an angle /~=sitl'(l/M). The Mach number on the plate OC having angle of inclination p= Boc=a, i.e. on the upper surface, is determined by woc = poc + o ~ . (7.4.2) The value of local Mach number Mot obtained helps in determining the Mach angle ,uoc=sinl(l/Moc). A graphic solution of the PrandtlMeyer flow is shown in Fig. 7.4.2, b. The coordinate of the point G' of the intersection of the epicycloid with the line O'G' parallel to the plane OC determines the relative velocity AOC of a disturbed flow near plane OC. In a physical plane the point G of the intersection of the stream line with the characteristic OE corresponds to point G'.
Fig. 7.4.2. System of flowinteraction of obtuse angle (PrandtlMeyer flow): aphysical plane; bhodograph plane; csystem of smallperturbation flow; Iexpansion wave; 2epicycloid.
From the known Mach number Moc (LOC) pressure on the upper side the of the plate can be found out by applying formula (3.6.30)
At hypersonic speeds (M,% 1) the analysis of the PrandtlMeyer flow is simplified because in determining the function w we can use (5.3.41), which helps to determine local Mach number M directly. Replacing angles woe and o in (7.4.2) by their values from (5.3.41), we get ,
The corresponding pressure can be determined by formula (3.6.39). Let us study the PrandtlMeyer flow arising at hypersonic speeds in the case of small angles of flow inclination Poc=a. At very high Mach numbers M, the bundle of Mach lines emerging from point 0 will be very narrow. With fair accuracy it can be assumed that these Mach lines condense to one line across which the flow turns suddenly and an expansion of flow takes place. Hence this line can apparently be regarded as the 'expansion shock' behind which speed (Mach number) increases and pressure decreases. The angle ,, u of inclination of this line to the vector 7; is obtained if the analogy of a compression shock is followed and the calculation of this angle is carried out by formula (4.6.9) under the condition that the angle PC POC a in this = = formula is negative. Assuming that 0, = pb,, we find
E l  2216) +J+1  ~ 5 ) ~ K2' 4( where I DOC I = I a I is the absolute value of the angle of turn of flow (angle
of attack) and the quantity K2=(M,j?~~)2. Applying the formula (4.6.1 1') for calculation of the coefficient of pressure, and taking into account the sign of angle POC a, we get =
poc   
(7.4.5)
At low Mach number M, and small angles of attack Poc=a PrandtlMeyer's small perturbation flow comes into the picture near the inclined surface. The expression for sonic speed is valid for such a flow. Now if the formula for the local Mach ?umber is found from (7.1.2):
and if the value of a2in this formula is introduced from (7.1.27, we get
In accordance with this expression we can assume as a first approximation that MmM, in a small perturbation flow and hence the characteristic equation in the physical plane will be dy/dx= tan(P + p,). As the angle of flow inclination p is small and P e p , we get dyldxm +tan p,. Consequently the characteristics themselves represent the Mach lines inclined to the x axis by the angles rt p,. For a PrandtlMeyer flow we have the family of characteristics in the form of parallel lines inclined to the horizontal axis by an angle p, (see Fig. 7.4.2, c). The equations of characteristics for a plsne supersonic flow in the hodograph plane can be obtained from (5.3.21) and (5.3.22) in terms of finite
differences:
(AV/V) T tan PAD =0. For a small perturbation flow we may take AV= u, V= V, AB =/I, Hence (7.4.8') u/v = k I?/~'M%  1. Substituting the value of u in (6.1.5) from (7.4.8'), we get the coefficient of pressure, tan ,u= tan ,x,=(MZ,
p=
72/?/6&
1.
(7.4.9)
Since a suction flow is assumed for which j < O and it is assumed that angle /3 is an absolute quantity the minus sign should be taken in formula (7.4.9). Accordingly the coefficient of pressure on the upper surface of the plate inclined at a small angle of attack B =a will be pu=poc U = 2al.d~;
1.
(7.4.10)
Let us examine the lower side of the flat plate. The flowinteraction of this side (see Fig. 7.4.1) is accompanied by the formation of shock OE emerging from a point on the leading edge and hence by the compression of flow. It is necessary to use the formula (4.3.25) to determine the angle of inclination of shock &OE,where MI=N , and PC= a. From the value of BcoE 1 obtained the Mach number Mz =Mocl on the lower side can be found from (4.3.19) or (4.3.19'). In determining the flow character in the region behind point C on the trailing edge the following~conclusions be drawn: On the upper side of the can plate the Mach number Mocu before the shock CD is higher than the Mach number M, before the shock OE appearing at the lower side of the leading edge. If we assume that the flow behind a point C does not deviate from the undisturbed flow direction (stream line CF is parallel to the velocity vector 5?,)it is obvious that the losses in the shock on the upper side will be heavier Me=!. Then the pressure from the top of line CF seems and therefore M c F ~ < to be more than that from below. Discontinuity in pressure on the boundary surface cannot be maintained in a gas flow although the velocities can remain different. Therefore in actual conditions the direction of the stream line CF differs from the direction of an undisturbed velocity, i.e. downwash is formed behind the plate. From the physical conditions it must be clear that the line CF inclines toward the lower side. Meantime the flow behind shock C D turns by a small angle, which leads to reductions in pressure.
Experiments show that the downwash angle is small and therefore it can be said with fair accuracy that the flow direction coincides with the direction of an undisturbed flow at a point C . Accordingly the shock angle &CDon the upper side is determined by the formula (4.3.25) where MI =Mocu, Bc=a. The corresponding Mach number behind the shock Mz= M c ~ u determined is from (4.3.19) or (4.3.19') for the given values of O,=&CD, MI=Mocu and PC a. = ~, The PrandtlMeyer flow with Mach number M ~ Fwhich can be determined with the help of formula wc~l=wocl+a,takes place below the line CF in the region behind the trailing edge. The pressure pu=pocu on the upper side of the plate is determined by formula (7.4.3) and the corresponding pressure on the lower side pl=pocl is obtained from the relation (4.3.1 5), where p2 =PI, PI =pw, 0, = &OE and P~<Pu<prn If the length of the plate is L and its width is equal to unity the force due to normal pressure acting on the plate is given by F=L (prp,). Hence the lift force is Y= F cas a and the drag force is X= F sin a. The corresponding values of the lift and drag coefficients will be c,= Y/q, L and cx= X/q,L respectively. Introducing the coefficients of pressure on the upper and lower and =(pl p,)/q, respectively, we get sides in the form &, = (pup,)/q, the following expressions for the aerodynamic coefficients: cy=(Fl&) cos a; The force X arising during the supersonic flowinteraction of the plate and induced by the formation of shock waves and plain waves of disturbances is known as wave drag and the corresponding quantity cx is called the coefficient of wave drag. This drag is not zero even in the case of an ideal (inviscid) fluid. The aerodynamic efficiency factor of plate K =cy/cx= cot a is obviously the function solely of the angle of attack. As a consequence of the uniform pressure distribution over the plate surface the center of pressure is situated at the middle. Hence the moment of pressure forces about an axis at the leading edge will be M, =F. L/2 and the corresponding coefficient of moment will be
At hypersonic speeds the coefficient of pressure on the upper surface of the plate is approximately obtained from formula (7.4.6) and on the lower surface from formula (4.6.12). Taking this into account and assuming Pc=a in (4.6.12), we get the following expression for the difference in pressure coefficients (jIj,), which is also known as the coefficient of pressure drop:
259
Hence
Formulas (7.4.13) through (7.4.16) indicate the law of hypersonic similarity applicable to the flowinteraction of a thin plate. The basic content of this law is that the corresponding values of $a2, c,/a2 c,/a3 and mz/a2for a flat plate will be one and the same for equal values of K=M,a but independent of the quantities M, and a. The parameter K =M,a is known as the criterion of hypersonic similarity. From formulas (7.4.13) through (7.4.16) it follows that the relations for the coefficients of pressure, lift and moment are quadratic. For the coefficient of drag it is the cubic function of the angle of attack a. In the limiting case at K+w:
For a flat plate situated in a smallperturbation (linearized) flow the pressure coefficients are calculated by formula (7.4.9) in which p = a is considered. The minus sign in this formula determines the pressure coeflicient on the upper side and the plus sign that on the lower side. Hence the difference between the pressure coefficients with respect to angle of attack is ( ~  j i J / a = 4 / * / ~ ~1. Substituting in formulas (7.4.1 1) and (7.4.12), we get c y / a = 4 / * / ~ k 1; (7.4.20) (7.4.19)
In the given case the criterion of similarity happens to be the Mach number M,. The corresponding values of ;/a, c,,/a, c,/a2 and m,/a will be identical independent of the value of the angle of attack but keeping the value of M, constant. If the case of a smallperturbation flow at very high values of Mach numbers M, > 1 is taken formulas (7.4.19) and (7.4.22) can be expressed in the following way:
In this way the criterion of hypersonic similarity K = &a is valid even for a smallperturbation (linearized) flow with high Mach numbers. It is obvious that this kind of flow can occur only at very small angles of attack.
5. Supersonic Flow Past Airfoil of Arbitrary Form 5.1 Application of method of characteristics Let us study a supersonic flow past a sharp airfoil of arbitrary form (Fig. 7.5.1). The upper and lower contours of the profile are given by the equations yu=fu (x) and y 1 = 5 (x) respectively. Let us assume that the angle of attack is more than the angle Pouformed by the tangent to a profile contour on the upper side at a point 0 at the leading edge. The PrandtlMeyer flow comes into the picture at this point. The flow assumes the form of an expansion fan emerging from point 0 as from the source of disturbances and follows the direction of the tangent to the contour at this point. Beyond point 0 the flow continues to expand along the contour. Consequently the flow past the airfoil can be considered as the successive summation of PrandtlMeyer flows. As the flow turns at the points on the contour by an infinitesimally small angle separate Mach lines will emerge from these points in place of the fans of expansion lines (see Pig. 7.5.1).
Iexpansion
fan; 2Mach
line.
The velocity at a point 0 is found with the help of formula (7.4.2), which can be written in the form where PO,, t a r 1(dyu/dx)o. The corresponding Mach number MOis obtained = from Table 5.3.1 for the given value of oo. At a neighboring point C situated at a small distance from the edge the
velocity of the gas is calculated as for a PrandtlMeyer flow with the help of the formula
Adding the left and right sides of equations (7.5.1) and (7.5.2), we get the relation where Qc= tanb1(dyU/dx)c. Hence, in accordance with (7.5.3) we have for any arbitrary point N on the contour
) where angle Q ~ = t a n  l( d y u l d ~ is~calculated taking into account its sign. For the front part of the contour the signs of angles will be positive and for the rear part negative. We assume that the flow at a point B, as in the case of the flowinteraction of a flat plate, approximately maintains the direction of an undisturbed ~ flow. Moving now at a speed corresponding to M B near the contour, the flow turns at point B and shock BE emerges therefrom. The angle of inclination B c ~ = the shock and the parameters behind it are worked out with the help of of the corresponding formulas of the theory of shocks for the known values of angle of attack a, Mach number M B and the contour angle at point B on ~ the upper side. The parameters on the upper surface of an airfoil (pressure, velocity, etc.) are determined by the relations for an isentropic gas flow from the known value of the local Mach number. If the angle of attack is equal to the angle PO,,we have the limiting case of the PrandtlMeyer flow at a point 0 for which the Mach number at this point Mor =M,. The formula (7.5.4) can be rewritten in the form
Calculation of flowinteraction on the lower side of the profile (Fig. 7.5.2) begins with the determination of gas parameters at point 0 immediately behind the shock. For this purpose the angle of inclination of shock BCo is I and PC =a + Pol. The Mach calculated on the basis of the values M I =% number Mol =Mz at a point 0is found from (4.3.19) or (4.3.19'). This Mach number will be assumed to be constant in the infinitesimally small region around point 0 on the segment of the contour in the form of a straight element OD. The straight element OJ of the oblique shock corresponds to this segment OD. Its length is determined as the distance from point 0 to point J, which lies at the intersection of the shock and the characteristic of the first family emerging from point D. The flow behind the rectilinear shock will be irrotational. Therefore the part of the contour behind point D is in the isentropic flow. To determine the
262
AERODYNAMICS
velocity of thiskind of flow at apoint F we use equation (7.4.1) from which we find w==wD(PD PF), where CUD is the value of angle cu worked out by formula (5.3.31) for a value of M on the segment OD of a contour. The values of angles and pF are determined taking into account the sign (in the present case the angles JYD and /3p on the front part of the contour are negative). !4 I
Fig. 7.5.2. Supersonic flowinteraction on lower surface of profile with formation of compression shock: Istraight portion of contour of body; 11curved portion of contour of body; 111curvilinear part of compression shock; IVlinear part of compression shock.
The flow around the contour segment DF can be considered as a PrandtlMeyer flow. Therefore the line of disturbance F13 emerges from point F as from the source of disturbance. It crosses the extension of a straight shock at point 3 and bends it. As a result the actual direction of the shock will be determined by the point of intersection 3' of the shock and the characteristic. The bending of the shock downstream will be governed by the interaction between it and the characteristics emerging from the points G, H, K, etc. The flow becomes rotational due to the shock bending. To calculate this it is necessary to apply the relations of a nonisentropic plane flow to the characteristics. The characteristic JH of the second family happens to be the boundary of this rotational flow. This characteristic will be gradually built up in the form of a broken line from the known values of Mach number M and angle p along the rectilinear characteristics of the second family emerging from points F, G, etc. on the contour. The velocity at point H on a contour lying at the same time on the characteristic JH is obtained from the expression w ~ = o D  ( ~ DpH). At point K, in the neighborhood of point H, the flow parameters are worked out from the equations of characteristics which take into account the rotational character of the flow behind a shock. To determine the velocity at this point it is enough here to know the speed and its direction (angle P) at point 7, which is situated near point K on the
element 7K of ,the characteristic of the second family. To determine this velocity we have to work out the curvature of the shock behind point J o n the segment J3' and find out the parameters on the shock at point 3'. The following data can be used for this purpose: Mach number M = M I =MF and angle P = j?p for the direction of flow at point 1 and the parameters M ~ = M J , on the segment at point J. Applying the formula (5.4.46) and assuming that E = O , we get the following expressions for the change in angle B along the characteristic 13:
where
AS (S3 S i ) cos (Pi + p i ) = An (x3xl) sin ,UI 7
C1=
The derivative (dw/dj?)j is calculated from (5.4.39) for the values of the corresponding parameters at point I; the angles wl and c o ~ obtained from are (5.3.31) for the Mach numbers M at points 1 and Jrespectively. Point 3 in Fig. 7.5.2 lies at the point of intersection of a rectilinear shock and the characteristic 13. Consequently its coordinates are found from the simultaneous solution of the equations:
~ 3  ~ 1 = ( ~ 3  ~tan(Pl+pl); 1 )
(7.5.6)
The intersection of the characteristic and the shock will be different (at point 3') because of the shock bending. The new angle of inclination 0cJ31of the shock on the segment 53' is determined from the angle of flow inclination behind the shock Pc=P3'=APl+P1 and the Mach number M3'. This M3' is calculated with the help of an expression w3'=Awl +col in which Awl is found from (5.4.38):
Hence the equation of the element J3' of a shock will be y,, yJ = (x,  X I ) tan (Bc~3,a). a result of the flow turning at point 3' the characteristic As changes its direction on the segment 13'. The equation of characteristics on this segment will take the form y3, Y I= ( ~ 3   X I tan (P3' + p 3 , ) . Solving 1 ) these two equations simultaneously7we get the coordinates x3,, y3, for point 3'. Let us examine a point 5 situated at the intersection of the elements of characteristics 25 and 3'5 of the first and second family respectively. The coordinates of this point ys and xs can be determined from the solution of the equations for the elements of the corresponding characteristics:
264 AERODYNAMICS Y S ~ 2 = (xs xz) tan (Q2 p2) ; Y Sy3# =( X S X Y ) tan (831p ~ ) . 
(7.5.8)
The change in the flowdirection during transition from point 2 to point 5 along the element 25 of the characteristic can be obtained from equation (5.4.22) with e =O:
3 . 1 is
The values of c2 and t3' are determined by (5.4.15). The Mach number M at a point 5 is found from (5.4.23) for the given AP2. Similarly the velocity at a point 6, lying at the intersection of the elements of characteristics H6 and 56 of the first and second family respectively, is determined for the known values of the gas parameters at point 5 and point H on the wall. Here the coordinates XH,y~ of point H are obtained from the solution of the contour equation Y H = ~ H ( X H )and the equation yHy2=(xHx2) x tan (p2 p2) for the element 26 of a characteristic. Let us take an arbitrary point 7 with coordinates y7, x7 on the element H6 of a characteristic. The flow parameters at this point are determined bjr the ~ interpolation process. For example, angle P7 = P 6 ( P 6 P H ) ( 7)/(6H). Similarly the Mach number M7 and the corresponding Mach angle p can be 7 found. Point 7 has been selected in such a way that the element 7K of a characteristic drawn from this point at an angle j17 p intersects the contour 7 at point K lying at a small distance from point H. The coordinates of this point y ~ XK are obtained from the solution of the equations yKY7 =(XKx7) , x tan (/37p7) and y~ =~ H ( X for the element of a characteristic and for a K) contour respectively. As the flow at point K is rotational the calculation of velocity at this point should be carried out with the help of (5.4.27). Assuming E = 0 in this equation, we get
where A07 =0~ 0 7 ; AP7 = fig  P7 and the angle PK= t a n  ' ( d ~ ~ / d xThe. )~ parameter t7 is determined from (5.4.28) for the given values of p and P7 7 and the entropy gradient is calculated by formula (5.4.29): A s (s7  SK)C s (P7 p7) O An  . (XKx7)sin p 7 '
The entropy Sg at point K is found by calculation of its value at a point 0 immediately behind the shock. The entropy ST at a point 7 is calculated by interpolation from its values at points H and 6 . Similarly the velocity field in the region between the characteristic DJ, the contour of profile and the curved shock 53'4' is determined by means of successive solution of each of the three problems examined in Chapter 5 Section 4. The form of this shock is determined step by step as a broken line and the given region of flow is filled up with a network of characteristics. At the nodal points of the net of characteristics pressure can be determined by using formula (3.6.28) for a given Mach number. In this formula stagnation pressure po=p; is calculated by interpolation from formula (5.4.19). The pressure at the points on a contour is found for the corresponding Mach numbers and the stagnation pressure po=p; calculated from (4.3.22) for angle 8,o and a given Mach number M,. Similarly we can determine the velocity, density and pressure at points on the lower side of a contour including point B at the leading edge. The flow calculation on the lower side of the airfoil is simplified in that case when the characteristic of a second family from point J does not intersect the contour, so that the flow can be considered isentropic (Fig. 7.5.3). For calculation of the Mach number at an arbitrary point L on a contour we use equation (7.5.1): tan'(dy~/dx)~, taking the sign Here angle PLisdetermined by the formulaP~= into account: it will be negative over the front part of a contour and positive near the leading edge; the angle will be negative. The form of a shock is determined by the angle Bc of its inclination at points 3' and 4'. These points are situated at the intersection of the shock and the rectilinear characteristics from points F G. The shock angle 0, at point 3' is approximately determined , from (4.3.25) for given values of P3' = PF and M,. Similarly we can find the angle of inclination of the shock at a point 4'.
Fig. 7.5.3. Supersonic isentropic flowinteraction of curvilinear sharp profile: 1contour of body in flow; 2characteristic of second family.
,The calculation of flowinteraction at the angles of attack a 2 pol has been examined for the case where there exists a suction flow all over the upper side of an airfoil. If this condition is not satisfied (a< Pol) the gas flow will be compressed over the upper contour near the leading edge and the shock will be formed there. Hence the calculation of this kind of flowinteraction should be carried out in the same way as on the lower surface.
5.2 Hypersonic flow over thin profile If a thin profile with sharp edges (see Fig. 7.5.1) is placed in a hypersonic flow at such small angles of attack that the local Mach numbers at all points of its upper and lower surfaces considerably exceed unity and if in addition the conditions of isentropic flow are maintained on the lower contour (see Fig. 7.5.3) the simplified relations of characteristics can be applied for the calculation of perturbation velocity. If angle of attack a Po,, then to calculate Mach number M at any arbitrary point N on the upper contour it is necessary to use, in accordance with (7.4.4), the formula
As there exists a suction flow all over the upper contour the Mach number MN >M,. The flowinteraction on the lower surface is accompanied by the formation of a shock and hence by the compression of the gas flow. The coefficient of pressure at a point 0 immediately behind the shock is calculated with the help of formula (4.6.12) in the following way:
where K =M, (a Pol) and the angle POIis taken with the minus sign. From the formula $=2eco (aPo,) it is possible to calculate the angle of shock inclination Bco and then the Mach number at point 0 can be found for the given value of K, =M,Oco by using formula (4.6.16):
The flow behind point 0 will be a suction flow, so to determine the velocity at an arbitrary point L we can use the relation
An approximate calculation of the hypersonic flowinteraction of a thin profile can be carried out with the help of the method of a local plate. According to this method the flow at any arbitrary point on a contour is worked
267
out from the corresponding formulas for a flat plate with the assumption that a plate is placed in a flow with a Mach number M, at an angle of attack equal to the angle between the velocity vector Qndthe tangent to a contour at the point in question. In this way the coefficient of pressure at some point N on the upper side of the contour will be, according to (7.4.6):
vb,
for the arbitrary point L on the lower surface of the airfoil. Here In formula (7.5.16) the angle BN has a plus sign for the front part of the contour and a minus sign near the trailing edge. In formula (7.5.17) the angle /ILhas a minus sign near the leading edge and a plus sign near the trailing edge. In the limit of K+oo : (7.5.18) (7.5.19) At zero angle of attack formulas (7.5.16), (7.5.17) and (7.5.19) take the following form:
PN = 0;
268 AERODYNAMICS
point N on the upper side according to (7.4.9), will be PNF=2 (a @,)I JM", 1 and at some point L on the lower side it will be PL=2(a8~/d~~l. For zero angle of attack we have
= 2pN/
Y
J M 1;  ~
1.
PL=2pL/4M2
5.4 Aerodynamic coefficients Formulas (1.3.2) and (1.3.3) can be used to determine aerodynamic pressure forces by transferring them to the fixed axes x, y (see Fig. 7.5.1). Formula (1.3.2) gives the axial force R for the airfoil and formula (1.3.3) the normal force h? R=q,So
(S)
 dS ~~os(P~x). So'
Taking SO be 1 as a characteristic area and recalling that dS = dl. 1 (b is the = chord of the profile, dl is the element of a contour), we get the following expressions for aerodynamic coefficients:
CR = Rlqoo o = S
CN =N/qws
o=
cos (pny)df, where df=dl/b and the integrals are worked out along the contour of an airfoil. Let us introduce the substitution
A
A
CC
=d%
where dx=dx/b. Now, coming to the usual integrals from the curvilinear integrals, we get:
where and i,, are the pressure coefficients on the upper and lower sides of the profile respectively. Applying the formula for transformation [see formula (1.2.3) and Table 1.2.11, we get aerodynamic coefficients in the velocity system of coordinates (Fig. 7.5.4):
cx=cR cos a + c sin a ; ~ c,=cN cos a  C R sin a.
] (7.5.24)
I (1.d~jyd~) q, Sob
,
Fig. 7.5.5. Scheme for determination BY analogy with (1.3.6) we get the of moment of forces for profile. formula for the coefficient of moment past the leading edge of the profile due to pressure forces (Fig. 7.5.5):
m,=
4, Sob
Mz =
where y=y/b. We will determine the pressure coefficient of the profile under the condition that the center of pressure lies on the profile chord. If its coordinate is xc.p,
1
~($P.)z~F+I [ ~ r n ( $ ) ~  ? ~ ? ~ ( g ) J d ;
c q = = b
Xc.,
0
f
0
. (7.5.26)
(5&) d ~
As an example let us find the relations of the aerodynamic coefficients for a wedgeshaped airfoil (Fig. 7.5.6). Recalling that (dyldx)~ tan PI for the =
the pressure coefficients $ and $ are constant, we get from (7.5.22) and
CR=
mtan/?r+&tanjlu;  CN=PI~U.
(7.5.27) (7.5.28)
Here angle 81 in (7.5.27) should be considered as negative. The coefficient of moment can be found from (7.5.25):
Fig. 7.5.6. Wedgeshaped profile.
At zero angle of attack the coefficients cRand CN determine respectively the coefficient of drag cx =CR and the coefficient of lift c, = CN. Two possible cases must be taken into account in calculating these coefficients. In the first case the angle of attack a >Bu and so there arises a PrandtlMeyer expansion flow on the upper side: hence &<0. On the lower side of the airfoil we find the compression of flow (c>O), which is analyzed by the theory of shock waves for angle of inclination =a The relations for calculation of the aerodynamic coefficients can be written in the form:
+ jar].
c,p=<an
(7.5.30) (7.5.31)
For a symmetrical wedge we have tan the angle of attack a =Bu, pu=0. For small wedge angles tan PEP; then
BU=ltan
in these formulas. If
For a hypersonic flow over a thin wedge formulas (7.4.6) and (4.6.12) can be used to calculate the pressure coefficients. The pressure coefficient on the lower side, using (4.6.12), will be
and on the upper side its absolute value can be found with the help of (7.4.6):
where K I = M = ( ~ + ( BI )I; K u = M , ( a  P u ) . Substituting the values of 6 and ~ & l from (7.5.33) and (7.5.34) in (7.5.301), (7.5.31) and (7.5.32'), we get the coeficient of axial and normal pressure forces and also the coefficient of moment due to pressure forces. If a thin wedge is placed in a flow not having high velocity the pressure coefficient is determined by formula (7.4.9), according to which
Substituting the values of& and pu from (7.5.35) and (7.5.36) in (7.5.30'), (7.5.31) and (7.5.32'), we get the coefficients of axial and normal forces and also the coefficient of moment due to pressure forces. Let us examine the second case of flowinteraction with a < Du. Now the compression of flow will be observed on the lower as well as on the upper side of the airfoil; consequently >0 and & >O. To determine the coeficients CR,CN and mz we have to apply formulas (7.5.27) through (7.5.29). For the arbitrary values of 8 and /la the pressure coefficients& and 1 will be obtained from the theory of shocks for angles of inclination /lc=a+ 1/111 and PC=Pu  a respectively. At hypersonic speeds the pressure coefficient on the lower side is calculated by the formula (7.5.33). On the upper side it will be found from the similar relation
where Ku =M,(P,  a). At low supersonic speeds formula (7.5.35) can be used. In place of (7.5.36) we can apply the relation pu=2 ( B ~  ~ ) / ( M " ,1. (7.5.38) For a symmetrical airfoil (IPII=Pu) and at zero angle of attack these formulas are simplified. At hypersonic speeds this case gives
where P = PI^= /lu, K =M&. According to the relations (7.5.27) through (7.5.29),
and hence
Zhukovskii's formula: We have seen from Chapter 6 Section 3 that the lift force acting on a profile in an incompressible flow is determined by Zhukovskii's formula (6.3.22). Experiments showed that this formula could be extended to the case of the flowinteraction of thin airfoils in a lowperturbation (linearized) gas flow with high speed. Let us examine this for the case of a flat plate at a low angle of attack in a supersonic tow. If the velocity of a disturbed flow over the upper side of a plate is equal to V +u and on the lower , side V u the circulation about the contour surrounding this plate will be , According to (6.1.5) the additional velocity u=1~,$21, and hence Using formula (7.5.31) in which it is assumed that we get the expression
As an example of the application of aerodynamic characteristics of an airfoil we will take up the analysis of the flowinteraction of a swept wing of infinite span in an inviscid gas flow. The scheme of this system is shown in Fig. 7.6.1. If IC is an angle of sweep this flowinteraction is characterized by the velocity component V,,=V, cos IC normal to the leading edge. Here it is assumed that the other velocity component V = V sin IC,parallel to this , , edge, does not call for any change in pressure distribution. Consequently the velocity field in each plane perpendicular to the leading edge will remain one
and the same. This indicates that the coefficients of pressure, lift and drag acting on a swept wing of infinite span could be determined from the values of these coefficients corresponding to the wing section (airfoil) in the plane perpendicular to its leading edge. The velocity and Mach number on the basis of which these coefficients are calculated will be V = V cos K and M,, = , , M, cos K respectively. Let us assume that in an undisturbed flow with supersonic speed (V, > a,, M, > 1) the sweep angle satisfies an inequality K >n/2  p,, according to , which cos K < sin p, and hence V < a, = V sin p,, i.e. the velocity component nor, mal to the leading edge will be subsonic. So M , ,? the flowinteraction of the section of a swept wing in its very nature will be subsonic. In this case the swept edge is known as a subsonic edge. At increased velocities the normal component may exceed sonic speed (V,, >a, = V sin p,) so that we have K < n/2  p, and , cos u > sin p,. In that case the flowinteraction of the swept wing profile will be supersonic. Then the leading edge will be called a supersonic edge. Let us study the calculation of the super Fig. 7.6.1. Swept wing of infinite span in supersonic flow: sonic flowinteraction of a swept wing for IMach line for subsonic edge; both these cases. 2Mach line for supersonic Supersonic leading edge: The flowinteracedge. tion of this wing can be worked out by the formulas derived for a flat plate of infinite span under the condition that the undisturbed flow velocity is V,,= V , cos u >a, and the corresponding Mach number is M,, =M, cos u > 1. The angle of attack of a flat plate a! is connected with the given angle of attack of a swept wing by the obvious relation sin ar=sin alcos K. At small angles of attack it will be al= alcos K. Using formula (7.4.9) and replacing B and M, in it by ar=a/cos K and Mn, =Moocos K respectively, we get a relation for the pressure coefficient on a swept wing:
v
B
*
In this formula pressure coefficient is taken with respect to the velocity head q,=(1/2) kp,~:,. To obtain the value of pressure coefficient with respect to the velocity head of an undisturbed flow q,= (112) k p , ~ L we have  to use the formula p=p, cos2 K, according to which
p= + 2a cos K / ~ M "cosZK  1. ,
(7.6.1)
274 AERODYNAMICS In (7.6.1) the plus sign determines the pressure coefficient on the lower side of the wing and the minus sign that on the upper side. The coefficients of lift and wave drag can be found from the expressions (7.4.1 1). Considering small angles of attack, we get c, =4a cos I C / ~ M L C O SIC 1, ~ cx= 4a2cos I
C / ~ cos2 IC  1. M ~
(7.6.2) (7.6.3)
By substituting the value of 1c=0 for unswept wings in (7.6.2) and (7.6.3) it can be seen that the sweep increases wave drag coefficient c,. At the same time the lift coefficient c, also increases. They are connected by the relation c, =c,/a. It will be shown in Chapter 8 that the results relating to an infinite aspectratio wing with sweep can be used for calculation of the flowinteraction of the parts of finite swept wings having supersonic leading edges. Subsonic leading edge: In this case the flowinteraction of the wing sections, corresponding to the motion of a rectangular wing with Mach number Mn,< 1, must be examined with the help of the subsonic or transonic theory of profile. The drag will be determined by the laws of subsonic flows and the wave drag may arise only during supercritical flow past the wing (M,,> M,,S, in which case shocks appear on the wing surface. If Mn,< Mmcr shocks and wave drag do not arise. This discussion is related to a wing of infinite span. Wave losses are always present for finite wings because the velocity component V , sin IC has an effect on the wingtips. As a result the supersonic properties of a flow come into the picture and wave drag appears. For the study of this drag it is necessary to apply the threedimensional theory of supersonic flowinteraction. Suction force: As shown in Chapter 6 Section 3, a suction force arises at the leading edge of an airfoil in an incompressible fluid flow. A similar effect appears in the case of a subsonic flow of compressible gas over an airfoil. Here the sweep of the wing's leading edge will have an effect on the value of this suction force. For calculation of this force we apply the expresFig. 7.6.2. Suction force of swept wing. sion (6.3.25) which can be
275
extended, with the help of the appropriate transformations, to the more general case of the flowinteraction of a wing with a swept leading edge (Fig. 7.6.2). Let us study this transformation. The velocity component of an undisturbed inviscid flow, tangential to the leading edge of a swept wing, does not change the field of perturbation velocities. It remains the same as for a rectangular wing in a fluid flow with velocity V,,,=V, cos K. The forces acting on the wing also remain unchanged. As a result of this a suction force will act on an element dzo of the wing with a straight leading edge (in the system of coordinates zo, xo) where, in accordance with (6.3.28'),
From Fig. (7.6.2) it follows that TO =dT/cos IC, ~ Z =~ Z / C O SIC, uo =U/COS O xoXL.E.O =(X XL.E) I<. cos Substituting these values in (7.6.4), we get d ~ / d z = n ~ c ' d ltan2 IC , + where
c2=
(7.6.5)
lim [ u ~ ( x  x ~ . ~ ) ] . L.E Expressions (7.6.5) and (7.6.6) can be generalized for the case of compressible flows. For this purpose see Chapter 8 Section 2 and use the relation (8.2.4) connecting the geometrical characteristics of wings in compressible and incompressible flows. From these relations it follows that all linear dimensions along the x axis for a wing in a compressible flow are reduced by d l M& times compared to those in an incompressible flow; at the same time the wing thickness and the spanwise dimensions remain unchanged. Therefore
X+X
where index "in" shows the dimensions of a wing in an incompressible flow. From (8.2.4) it follows that
and, therefore the pressure coefficients in compressible and incompressible flows are related by
p=pin/ 41 M:. (7.6.9) Suction force T, by its physical nature, happens to be a force dependent on the action of normal pressure. It is determined from the condition TwaY at small angles of attack. The corresponding coefficient of suction drag will be given by ex,= T/(q,So) =ac,. As the wing dimensions along the y axis do not change as between compressible and incompressible flows the angles of attack for the two kinds of flow also remain unchanged, i.e. a=ain. Hence
or, according to (7.1.14),
CxT=CxTin/
 
"k.
(7.6.10)
Keeping in view the relations obtained, the following expressions can be written for the airfoil of a swept wing:
CXT
2 2 p, V , pWV, 2bdz= CxTin . 2 bin J l M2, dzin 
41M:
dT/dz =dTin/d~in.
The right side of equation (7.6.11), dealing with an incompressible flow, is determined by formulas (7.6.5) and (7.6.6):
c&=
in
lim
X . +X
[uin(xi,x ~ . ~ . ~ ~ ) ] .
L.E.in
Hence
277
LE .
7. Airfoil in Supersonic Gas Flow with Variable Heat Coefficients Let us consider the analysis of a flowinteraction process for a wing profile in a hightemperature supersonic flow of gas in which the variation of specific heats can be defined by balanced dissociation. We assume that the flow is potential about the curvilinear surface of the profile and that it represents by itself a suction flow described by equation (5.3.27) for the calculation of an angle o of turn of flow from the initial value of Mach number M equal to unity. The integral on the right side of this equation can be determined beforehand by giving a series of values of enthalpy il < iofor selected fixed values of stagnation enthalpy io and entropy S. The calculations are carried out in the following way: The corresponding values of velocity should be worked out from the formula Vi= d2(i0 il) and the sonic speed ai could be worked out for the given values of ii and S from the table or graph of thermodynamic functions. Now from the values of ai and Vi obtained above it is possible to calculate pi= sin' (ai/Vi) and cot pi = d f1 ~. Among the values of Vi and ai we find certain values that happen to be the critical ones (Vi =ai =a*). The value of a* so obtained is the lower limit of the integral (5.3.27). The upper limit is any arbitrary quantity selected from the condition of Vi >a*. From the results of the numerical solution of this integral tables are composed [13] containing the data for the values of entropy S from lo4 to 1.9 x 104 m2/sec2. degree at an interval of A S = 103 m2/ sec2. degree and the maximum values of stagnation enthalpy io to 4 x 107 m2/ sec2 at an interval of Aio= lo6 m2/sec2. These are selected as the fixed parameters. The corresponding values of velocity V in m/sec, sonic speed a in m/ sec, Mach number M and turning angle coo are given in the tables for each pair of values of io and S depending on the selected values of enthalpy i with an interval of Ai= lo5 m2/sec2. With the help of these tables the problem of a dissociating gas flow around a sharpedged airfoil of curvilinear form under given conditions of undisturbed flow (V,, i S a) is solved in the following way: , , To determine the velocity on the upper side of the airfoil (see Fig. 7.5.1) with suction flow all over we must use formula (7.5.4). The value of owin (7.5.4) is found from the table [13] for the given values of enthalpy io=i,+ 0.5V; and entropy S,= S ( p , , T) After selecting an arbitrary point N at ,. which the inclination of the surface to the vector is determined by an angle (a j?), the angle o~can be determined by (7.5.4) and then velocity VN and the corresponding Mach number MN can be found from the same tables [13] for given io and S.
The calculation of flowidteraction with formation of shock on the lower side can be carried out in the following way: First we work out the parameters of a dissociating gas behind the shock near the leading edge point 0 (velocity Val, pressure pol, temperature To/, enthalpy iol, entropy SI). Then io = iol f 0.5 VZ0l is found from the values of ioland Vol and the angle wol is determined from the tables [13] for given values of io and S. The flow behind the shock will be rotational (nonisentropic) and the calculation of its parameters should, strictly speaking, be performed taking this peculiarity into account. But this of course makes the analysis complicated. However, the simpler method based on the application of formula (7.5.4) for an isentropic flow appears to be useful for evaluation of these parameters. For a given point L on the contour this formula will be written in the form
OL=UOI
 (POI PL), 
where POI,PL are the angles of inclination of tangents at point 0 on the leading edge and point L respectively. The velocity VL and Mach number ML are found on the basis of the value of w~ obtained from tables [13] for given values of io and S. The pressure, density and temperature at point L are determined from the iS diagram or from the appropriate tables in the form of the functions p(i,, S), p(iL, S), T(iL, S ) for iL= io0.5 v : . It is obvious that the calculation of the parameters of a dissociating gas over a flat plate is simplified. Let us consider the upper side of it. From the known values of V p,, T, we determine i,(p,, , T,) and S(p,, T,) and also , the stagnation enthalpy io= i,+O.5 VL. Next we find w from tables [13] for given values of io and S and then work out ooc=o,+a. The remaining parameters will be determined from the same table for this value of wo,. The calculation of flowinteraction on the lower side is related to the calculation of the dissociating gas parameters behind an oblique shock for a given angle of inclination of the plate and the known parameters of an'undisturbed flow (see Chapter 4 Sections 7 and 8). Here it is assumed that the parameters obtained for the conditions immediately behind the shock will be the flow parameters on the lower side of the plate.
1.1 Linearization of equation for a potential function Let us consider a thin, small cambered wing of arbitrary plan and finite span in a supersonic flow at a small angle of attack. The disturbances created in the flow by such a wing will be small. The linearized theory used to study a smallperturbation flow around a thin airfoil (see Chapter 6 Section 2) can be applied to the flowinteraction problem of this kind of wing. The velocity conditions for such flows were given in the form of (6.1.1). If a linearized threedimensional gas flow is considered these conditions must be completed by adding the terms for the velocity component along the z axis. Accordingly the following relations will be valid for a linearized threedimensional flow with small disturbances:
where u, v, w are components of the perturbation velocity along the x, y, z axes respectively. By virtue of the property of a linearized flow we have ugV,, vgV,, wgV,. (8.1.2)
These conditions help to linearize the equation of motion and equation of continuity and simplify the solution of the flowinteraction problem of a thin wing in a steady inviscid fluid flow. The equations of motion of this kind of flow in their general form are obtained from the system (3.1.17), where p =o, v,/at = a vy/at= v,/at =o:
280
AERODYNAMICS
'The continuity equation is assumed in the form of (2.4.4). Expanding the partial derivatives, we get
As shown in Chapter 5 Section 1, the equation of motion and the continuity equation can be combined into one equation relating the velocity components. Performing the transformations as in Chapter 5 Section 1, we get this equation in the following form:
Keeping in view the relation (2.3.2) for the potential function and the equality condition of partial cross derivatives (a2p/a.@y= a2p/ayax, etc.), the equation of velocity potential from (8.1.5) will be as follows:
These equations (8.1.5) and (8.1.6) are the basic differential equations of gas dynamics for threedimensional steady gas flows. The first corresponds to the more general case of rotational (not potential) gas motion and the second is used only for the study of irrotational (potential) flows. As the flow around thin wings at small angles of attack is potential the equation of velocity potential (8.1.6) can be applied to the study of this flowinteraction. To linearize equation (8.1.6), which is a nonlinear differential equation, we introduce into it the expression for sonic velocity (7.1.2') and the quantities
p=pm+pf. Analyzing the equation so obtained from the viewpoint of the order of smallness of the terms, as we did in Chapter 6 Section 2 for a smallperturbation plane gas flow: we get the linearized equation for the velocity potential of threedimensional perturbation components in the following way:
1.2 Boundary conditions Study of the flowinteraction of a thin wing of finite aspect ratio leads to the solution of the linearized equation (8.1.7) in terms of secondorder partial derivatives of the velocity potential p' for given boundary conditions. Let us examine these boundary conditions. 1. The wing in a linearized flow (Fig. 8.1.1) induces disturbances which are concentrated in the region of waves. This region is bounded by the surface representing the Mach cones with their apexes at the points on theleading edge with the angle ,u,=sinI (l/Mw). The boundary condition to be satisfied by the solution of equation (8.1.7) for function q' can be written as
By this condition the perturbation velocities are equal to zero on the surface of the wave (denoted by C ) or outside this surface.
'right
2. The solution for an additional potential q' must satisfy the boundary condition of an unseparated flow over wing surface S so that the normal component of velocity at every point on the wing is equal to zero, i.e.
A ( v ~ a~ = ( cos (nx) += cos (ny) , ~ ) ~ a~ ~ a~ a~ cos (nz) =0. * +
a2
Here
The directional cosines of the external normal to the surface are found from the formulas af analytical geometry:
cos (nx) =  A (aflax), cos (nj) = A, cos )nz) =  A (aflaz), where and the function f is determined by the nature of the wing surface [function y= f (x, z) for an upper surface, function y=fr (x, z) for a lower surface]. , The linearized (smallperturbation) character of the flow is realized under the condition that the wing is thin and hence af/ax< 1, aflaz< 1. Accordingly cos (nx) =  aflax, cos (ny) = 1, cos (nz) =  aflaz. During the flowinteraction of a thin wing the following conditions are also fulfilled:
A A A
A
(8.1.10)
Taking this into account, equation (8.1.9) can be written in the form
3. The flowinteraction may be accompanied by the appearance of a lift force whose resultant value is obtained by integrating the values of the elemental lift force [for wing elements having width dz and lengths equal to the chords b (z) (see Fig. 8.1. I)]. According to the expression (6.1.8) the coefficient of lift for an elemental surface, assuming circulation T(z) at the section z equal to
will be
Hence the circulation at a given section is From the correlating equation (8.1.13) (see Chapter 6 Section 4) it is clear that the circulation changes as we move to a neighboring section with a different lift coefficient. This variation is given by d r (z) = (drldz) dz =0.5 V, (dldz) (cryb)dz. (8.1.14)
According to the vortex model of a wing discussed in Chapter 6 Section 4 the elementary bound vortex of the section in question must pass through the contour surrounding the neighboring section. This vortex undergoes a turn and leaves the trailing edges in the form of a pair of elementary free vortices forming a vortex sheet behind the wing (see Fig. 8.1.1). In the case of a thin wing at a low angle of attack the width of this sheet may be taken as the wingspan and the free vortices may be directed along the direction of the undisturbed flow. The following boundary conditions on a vortex sheet may be established from the physical considerations. The normal velocity component of a particle vn= apr/an must remain continuous on it. Since the direction of the normal over the vortex sheet differs negligibly from the direction of axis Oy the derivative av'lan can be replaced by the quantity aq'lay. Consequently the condition may be written as (apl/a~)y=+o( a ~ f / a ~ ) y =  o y = (8.1.15)
where the left side corresponds to velocity V, immediately above the vortex sheet (y= +O) and the right side to that below it (y= 0). The condition (8.1.15) represents continuity of the function apf/ay across the vortex sheet. In addition the dynamic condition of continuity of pressure must be satisfied on the vortex sheet. From (6.1.5) we get the relation ( ~ P ~ I ~ X =~ a p+ ~ a ~ ) , =  ~ , ) (= tO (8.1.16) indicating the continuity of the derivative apr/ax during transition through the vortex sheet. Let us examine the flowinteraction of a wing of symmetrical profile (jVuyr) at zero angle of attack. In this case there is no lift force and hence = no vortex sheet. The vertical components of velocity on the upper and lower sides are equal in magnitude and opposite in sign due to the symmetry of the wing, i.e. v (x, +y, z) =  v (x, y, z). The component v = O in the plane xOz outside the wing. Hence a p ~ =0. a ~ ~ (8.1.17) Next let us assume that a wing of zero thickness and the same plan given by the strrface equation y = f (x, z) is placed in the flow at a small angle of attack. It follows from (8.1.12) that the vertical components of velocity V, =&'lay on the upper and lower sides of the wing at the corresponding points are equal. In view of the rather small angle of attack the above condition can be used for the plane y=O. At the same time this condition can be extended to the vortex sheet behind the wing, which is treated as an extension of the vortices in the plane Y =0. Therefore the components of velocity V , at the points situated symmetrically about this plane are equal, i.e. ay1 (xYYY z)iay = avr/(x, +Y, z)/ay.
Consequently the additional potential p' is an odd function of coordinate j~, i.e. Therefore, the derivative apf/ax on the lower side of the vortex sheet is equal to the value of apl/ax on the upper side. However, the equality of the derivatives ap'/ax was established from the condition of continuity of pressure. The above equalities may be simultaneously fulfilled if and only if on the vortex sheet. 4. To establish the last boundary condition we take the disturbed regions Srighly Sleft (see Fig. 8.1.1) representing the parts of the plane y =0 divided by the Mach wave surface situated outside the wing and vortex sheet. The flow is continuous over these parts of the plane y=O in the limits of the Mach wave zone. Therefore the potential p' also turns out to be a continuous function. At the same time, taking into account that function p' is an odd function according to (8.1.18), it is necessary to take
and that of a wing of zero thickness by an equation of mean lines of the airfoil In this way the resultant potential for a given wing will be
pf=p; +pi.
(8.1.23)
The flow about wing 2 at af 0 can be represented as the flow about a wing 3 at zero angle of attack with the surface equation y=0.5 (f,+Ji) and an addi
tional flow superposed on this taking place about a wing 4 in the form of a flat plate with the same chord as the original wing placed at an angle a (see Fig. 8.1.2).
gtven profile
mean line
I
mean line
2)
chord
Fig. 8.1.2. System of linearized supersonic flowinteraction of wing of finite thickness at a given angle of attack: Iwing at a=O with symmetrical profile and given thickness distribution; 2wing at a#O with zero thickness (wing profile coincides with mean line); 3wing at a=O with zero thickness (wing profile coincides with mean line); 4wing a t a#O with zero thickness (wing profile in form of flat plate of same chord).
According to this system of dividing the flow about a wing into two additional components of flow the resultant potential for a given wing will be
By analogy with this value of the velocity potential the distribution of pressure coefficientmay be given by
and then the drag force and lift can be found. From (8.1.24) and (8.1.25) it can be seen that the resultant drag force of a given wing can be obtained from the forces for wings 1,3 and 4, i.e.
Introducing the notation for the sum of two components X3 and X4 in the form Xi= X3 X4 and writing the above equation of drag force in terms of aerodynamic drag coefficient, we have
or,
From (8.1.27) it follows that the coefficient of drag of a wing consists of the coefficient of drag c , ~of a symmetrical wing at cy =O and the additional drag coefficient cxi related to the lift force calculated for a wing of zero thickness at c,#O. This coefficient cxi is further composed of the coefficient of induced wave drag calculated for a case where induced vortices are absent and the complementary coefficient of induced vortex drag due to the finite span and the consequent formation of a vortex sheet behind the trailing edge of the wing. By analogy with the expression (8.1.26) for drag we can write, in a general form, the relation determining the resultant lift force on a wing Y= YI+ Y + 3 Y4. It can be seen from Fig. 8.1.2 that wing 1 having a symmetrical profile creates no lift force at zero angle of attack, i.e. Y I=O. Hence the resultant lift force on a wing is
Y = Y3+ Y 4
and the corresponding coefficient of this force is
(8.1.28)
In this way, according to the approximate linearized theory of flowinteraction, the wing thickness will have no effect on lift force. Wing 3 gives constant lift force independent of the angle of attack and represents the value of this force at zero angle of attack for a given camber of wing. The lift force related to the angle of attack is provided by wing 4 and hence depends on the wing plan. The main problem of the aerodynamics of finite wings in a smallperturbation supersonic flow lies in finding the pressure distribution resulting in forces and the corresponding aerodynamic coefficients in terms of their individual components from formulas (8.1.27), (8.1.29). This problem also forms part of the basic material of the present chapter. 1.4 Characteristics of supersonic flowinteraction of wings  When determining the aerodynamic characteristics of wings it is necessary to take into account the peculiarities of supersonic flowinteractions. The peculiarities are based on the specific property of supersonic flows whereby the disturbances spread only downstream and within the boundaries of the cone of disturbances (Mach cone) with the angle p,=sinl(l/M,) at the apex. Let us examine the flowinteraction of a thin wing of arbitrary plan in a supersonic flow (Fig. 8.1.3). Point 0 at the leading edge happens to be the source of disturbances spreading downstream in the limits of the Mach cone. Here the Mach lines O F and OG can be taken in front of the leading edge
(Fig. 8.1.3, a) as well as behind it (Fig. 8.1.3, b). The position of Mach lines for a given form bf wing depends on the Mach number M,. In the first case the Mach number M, is less than that in the second case and the angle of disturbance ji, > (42) rc (rcangle of sweep back). The component of velocity normal to the leading edge is V,,=V, cos rc. As cos K <sin p,= 1/M, and V =a,M, it is obvious that the normal component V,, is less than the , speed of sound. The modon of a gas in the region near the leading edge of a swept wing for this case was studied in Chapter 7 Section 6 . This motion is similar to the subsonic flowinteraction of an airfoil which is characterized by interaction between the upper and lower surfaces occurring through the leading edge. This kind of edge is called a subsonic leading edge (see Fig. 8.1.3, a).
0 '
Fig. 8.1.3. Scheme of supersonic flowinteraction of wing: awing with subsonic edges; bwing with supersonic edges.
With an increase in speed the zone of propagation of disturbances diminishes and the Mach lines remain behind the leading edges. In this case, as shown in Fig. 8.1.3, b, the normal velocity component becomes higher than sonic velocity. In reality, it can be seen from Fig. 8.1.3, b that the angle of disturbances 11, <  K. Hence sin u = 1/M, c cos IC and therefore Vn,= , , 2 V , cos rc > a,. This kind of leading edge is called a supersonic leading edge. The flowinteraction of a wing in the region of this leading edge is supersonic in nature. The peculiarity of this flow is the absence of interaction between the upper and lower surfaces. If the Mach line coincides with the leading edge (rc= w/2p,) the edge will be a sonic one. It is obvious that the velocity component normal to the leading edge in this case is equal to the speed of sound. Let us introduce the parameter of sweep n =tan rclcot ,uC1,.For a supersonic leading edge cot ji, >tan rc and therefore n < 1. With subsonic and sonic leading edges we have n > 1 and n = 1 respectively because in the first case cot ji, < tan rc and in the second cot u =tan rc. ,,
T 'C
,By analogy with the leading edge the concept of subsonic, sonic and supersonic wing tips and trailing edges can be introduced. Wing tip CD, having an angle of inclination to the direction of undisturbed flow velocity less than the angle of disturbances (see Fig. 8.1.3, a), is called a subsonic tip. The component of velocity normal to the tip and equal to V,,= V sin ys in the given , case will be less than sonic velocity. In reality, since a,= V sin p, and p, > , ys we have V,,< a,. It is obvious that the parameter of sweep will be n > 1. The part of the wing surface with the subsonic tip lies in the region bounded by the cones of disturbance emerging from the points of discontinuity A and C of the contour. A crossflow of air across the wing tips will be observed due to the pressure of the subsonic normal component of velocity determining the flowinteraction over this part of the wing. This will result in a change in pressure distribution. These effects of wing tips on the flowinteraction of a wing are not found in general if the wing tips are supersonic, which will happen when ys >p, (see Fig. 8.1.3, b). In this case the normal velocity component V,,= V , sin y8 is higher than the sonic velocity a,= V sin p,. , Similar conclusions may be derived for the trailing edge of a wing. A subsonic trailing edge (y3 < p; V < a,) and a supersonic trailing edge (y3 > , n p,; V,, >a,) are shown in Figs. 8.1.3, a and 8.1.3, b respectively. From the above analysis a qualitative difference between the supersonic and subsonic flowinteractions of a wing can be explained. This difference enters the picture due to the different types of effects of wing tips and trailing edges on the flowinteraction over the wing surface. Whereas the wing tips and trailing edges in a supersonic flow do not affect the flow near the wing at all (see Fig. 8.1.3, b), or this effect is restricted to the part of the wing surface in the vicinity of these edges (see Fig. 8.1.3, a) the effect of wing tips and trailing edges in a subsonic flow will be felt all over the surface due to the possibility of propagation of disturbances downward as well as upward.
2. Method of Sources
The method of sources is used to solve the problem of the determination of the aerodynamic characteristics (PI, c , ~ ) of a thin wing of arbitrary plan and symmetrical profile in a smallperturbation flow at zero angle of attack
(cy=0).
.
The sources in an incompressible flow were discussed in Chapter 2 Section 9. The velocity potential of flow due to an incompressible point source situated at the origin of the coordinate system xn, y,, zn according to (2.9.14) will be
qn=  qn/4n J x i
+y," +Z; ,
(8.2.1)
where q, is volume flow from the source per unit time. In the method of sources we are not dealing with isolated point sources
but with sources continuously distributed over some part of a planeusually the coordinate plane xOz. Let dq, be an elementary volume flow rate of fluid created by sources on a small area do, = dt,d(, in plane xOz. Then the derivative dq,/do, = Q,, known as the intensity of distribution of sources, defines the strength of sources per unit area. If + v is the vertical component of velocity on an elemental area (the plus sign indicates that the fluid emerges from sources on the area and the minus sign shows downward source flow) it is obvious that the elemental volume flow rate dq, =2vda, and hence
For this elementary source the corresponding velocity potential will be (8.2.3) dvi =  Q, don/4z .x +y +zi . (i : Using this expression we can find the relation for the elementary potential of a source in a subsonic compressible flow. Let us examine equation (8.1.7) for this purpose and introduce new variables x,=x/dl M%, yn=y, zn=z. (8.2.4)
With the help of these variables equation (8.1.7) will be transformed into
which coincides with the continuity equation (2.4.8') for an incompressible flow. Hence the problem of a compressible perturbation flow in coordinates x, y, z can be reduced to the problem of an incompressible perturbation flow in coordinates x,, y,, z,. Here both systems of coordinates are related to equations (8.2.4). So it is possible to go from the potential (8.2.3) for an incompressible elementary source to the corresponding potential for a compressible subsonic source. Let us find the relation between the small area do, in plane xnOzn for an incompressible flow and the area do in the comparable plane for a compressible flow. Using equation (8.2.4) (by replacing xn with t and z, with ( and the expression do, =d<,d(, we find dc, =(dtd[) (1141 M%), from ) ' which, recalling that dtdc =do, we get the relation
do,=doldl M&. (8.2.6) Further we transform the expression for Q, appearing in (8.2.3). The component v n = aql/ay, or, according to (8.2.4), v, aql/ay. It follows from this that the component of velocity v in a compressible flow is equal to the component of velocity v, in an incompressible flow. That is, the densities of distribution of sources in compressible and incompressibleflows are identical:
290 AERODYNAMICS
Taking the relations obtained above into account, equation (8.2.3) can be transformed into the following form for a compressible flow: dp' =  ~ d a / 4 n d x (1 ML) (y2+z2). ~
(8.2.8)
Direct substitution of this indicates that the function p' is an integral of equation (8.2.5). Here it is immaterial whether the speeds are subsonic (M,< l) or supersonic (M, > 1). In the latter case the expression for an elementary potential may conveniently be written in the form dp' =  eda/4ndx2 af2(y2 zZ),
(8.2.9)
where aj2=ML  1. It can be seen from the expression (8.2.9) that at M, > 1 it has real values in the part of space where x2 af2(y2+z2). This indicates that the region of influence of sources, i.e. the region of perturbed flow witnessing interaction of these sources, lies inside the conical surface represented by the equation x2=af2(y2+z2). If a point is situated outside this surface, the sources will have no effect on that point. Here perturbation flow due to the given source is absent. Equation x2=af2(y2+z2) formally determines the surfaces of two coaxial cones (Fig. 8.2.1) with their apexes at the origin of coordinates. Hence the strength of a source Qda is used in creating perturbation flows inside these cones. In an actual case the supersonic disturbances propagate only downstream, i.e. only in one of the cones (righthand cone in Fig. 8.2.1). The perturbation flow in such a cone is determined by the potential having twice the value of that in (8.2.9) because the whole strength of the source, not just half its value, is felt in the flow enclosed in the cone of disturbance. According to this
In the previous case the elementary area da=d<dc with sources was placed at the origin of coordinates. If it is moved from the origin of coordinates to the point with coordinates x=<, z=c, equation (8.2.10) will be written as
During the study of the flowinteraction of a wing its surface is replaced by a system of distributed sources. To find the potential due to these sources at a point A ( x , y, z) (Fig. 8.2.2) it is necessary to integrate (8.2.1 1) over the region a in which at least part of the sources are situated. Each of these sources will have an effect at a point A (x, y, z) if this point lies inside the
cone of disturbances with its apex at the source. Thus the zone of action of sources (region of integration) is assumed to be in the region of intersection of the wing surface and the "inverted cone of disturbances" having its apex at a point A (x, y, 2).
Fig. 8.2.1. Regions of disturbed flow due to supersonic source; rightcone of disturbances (Mach cone) in real supersonic flow.
In the simpler case the point A and the source are placed, as shown in Fig. 8.2.2, in the same plane y=O. In this case the zone of action coincides with the region of intersection of the wing and the lines of disturbances emerging from a point M (x, 2). The region of integration lies on the wing and represents the intersection of the wing with the "inverted plane Mach wave" with its apex at the point A (x, 2).
After determining the region of integration a it is possible to find the resultant potential at a point A (x, y, 2):
The additional axial component of velocity can be found by taking the partial derivative of 9' in (8.2.12):
292
AERODYNAMICS
The pressure coefficient at the corresponding point obtained from this expression. Let us introduce the new coordinates
i= 2
(u/V,) can be
In the particular case of a point lying on the wing ( y l =0) the .additional potential
where
Q =2~,,=0=2 ( ~ ( D ' / ~ Y I ) , ~ = o
according to (8.2.7). The expressions obtained for the potential function help us to find the distribution of velocity and pressure over the surface of a thin wing for a given plan, shape of wing profile and Mach number M of undisturbed flow.
3. Delta Wing with Symmetrical Profile (a = 0 c, =0) ,
3.1 Semiwing with subsonic leading edge Let us study the flowinteraction of a semiwing with symmetrical airfoil in a supersonic flow at zero angle of attack. This semiwing represents a triangular surface with one of its sides lying on the x axis and the trailing edge
293
at infinity (this kind of semiwing is also known as a delta semiwing of infinite length) ( ~ i ~ . ' 8.3.1). If this surface has a subsonic leading edge the Mach line emerging from apex 0 will be situated ahead of this edge. The parameters of flowinteraction a t small angles of attack can be determined by replacing the lifting surface with a system of sources distributed in the plane y = 0. Consider an arbitrary point P on the surface and calculate the velocity potential at this point by adding the effects of sources lying in the region OAPB bounded by the leading edge and side edges O A and OB and the Mach lines AP and BP. The intensity of sources Q (5, [) is determined by formula (8.2.7) in which v =IV, from the condition of unseparated flow, where 1= dyldx is the slope of the wing surface. In this way
The velocity potential at point P is obtained from formula (8.2.12). Substituting Q by 21V, and taking y =0, we get 132
~t
11
0
o2'
(8.3.1)
where XP, ZP are coordinates of point P. This integral takes into account the effect of sources lying on the area o a t point P . This area o is equal to the region OAPB which can be represented in the form of the sum of the two segments OAPH and HPB. Accordingly the integral 9' in (8.3.1) can be written in the form of the sum of two integrals:
,
HPB
where
The integration in the section OAPH with respect to [ for each of the value of 5 = can be taken from g = &= tan K up to 5 =tD XP (ZP 5)a' = and the integration with respect to 5 can be taken from 0 to zp. On the section HPB it is necessary to carry out integration with respect to 5 for values of [=5 2 from 5 =<F =[ tan K up to t =<E= XP  (5 zp)a'. Integration with respect to 5 should be carried out from zp up to ZB =(XP+alzp)/(a' +tan K). Thus
xpa' (zp5)
rt="[[d~It
0
ZB
xpa' (zp5)
~ ( t y ~ * t + J d ~
zp
f ( t . ~ d t (8.3.4)
2: tan K
5 tan K
294 AERODYNAMICS
Using this expression and applying the main value of the integral, we get the following formula for potential function:
where the function IzP(I is taken with its absolute value. Calculating the partial derivative apl/ax, we get the component of additional velocity at a point P along the x axis:
dt d(xPl tan ~ ) 2  (zp ~ ) 2 a12
0
(8.3.7)
. . .+ 22,
In
[2 d(tan2
Substituting here the value of z~= (xp +a'zp)/(al + tan ,K), we get after simplification: a' (zp tan K XP) In U= 7cJtan2 K af2 aI2zp XP tan K + d(tan2 K af2)( ~ 2 , ar2z%) ' The expression under the square root sign in the denominator can be brought to the following form: ~ (tan2 K af2)(x$  af2z;) =(xp tan K  ar2Z P )at2(xp zp tan K ) ~ . Consequently
U=
Jp
3I
U=
(8.3.8)
For convenience of calculation we introduce angle B determined by the condition tan 0 =zp/xp and the angle of sharpness of the leading edge y =n/2  u (Fig. 8.3.2). We also introduce the notations
Let us now consider point N lying on a wing in the region between the Mach line OK' and the x axis (see Fig. 8.3.1) and calculate the velocity at this point induced by the sources distributed over the wing surface. Formula (8.3.1) for determining velocity potential can be used for this purpose. Taking into account that the effect of the source at a point N is restricted to the region a = OLJ, we get the expression
( ,= $'
1(<,C) d<dC, 51 f It
OLJ
where the function f (c, [) is determined by the relation (8.3.3). The integration with respect to c for each value of [=C3 should be carried out from c= ~ R = Ctan K to =&=xN ta' (ZN c) and the integration with respect to 5 should be done from 0 to ZJ. Thus
zJ 7C
x~+a'(z~O
J d[
0
J
5 tan K
f (535) d<,
(8.3.12)
where zJ = (XN afzN)/(tanIC +a'). On integrating and then using the main value of the integral we get
p'=  cosh1 ~c
0
XN5 tan K 4
This expression is similar to (8.3.6) with the difference that the upper limit of integration here is taken as the coordinate of a point ZJ. After finding the derivative ay'lax and carrying out the integration we get the relation (8.3.9') for an additional velocity component where it is necessary to take a<O because the coordinate z~ happens to be of negative value. It is possible to take the coordinate ZN as positive in calculations and then a > 0 If an . absolute value of tan IC is taken in this process it is possible to use equation (8.3.9) for determination of the induced velocity. In (8.3.9) the sign before a should be changed for this case. The relation will then have the following form:
The sources distributed over a wing also induce velocity in the region between the Mach line OK and the leading subsonic edge (Fig. 8.3.3). The value of this velocity at any point L is determined by the sources distributed
in the region OUG. The corresponding potential function can be obtained from the expression (8.3.12) in which ZJ and xN+at (zN() should be replaced by zu and xLa' (ZLl;) respectively where the lastnamed is the coordinate of a point R (see Fig. 8.3.3). Thus pt='5lddr n
0
1
f tan
K
f(<7i)d<r
(8.3.15)
After finding the derivative apr/ax and then integrating under the condition that o=zL tan ~ / x p > l get, as for (8.3.9), the following formula for we additional velocity component:
Fig. 8.3.4. Pressure field for triangular panel of wing with subsonic leading edge.
The pressure field for a semiwing of triangular plan with a subsonic leading edge is shown in Fig. 8.3.4. The coefficient of pressure along the Mach lines is equal to zero. The theoretical value of the pressure coefficient
on .the leading edge is equal to infinity. The pressure realized can be considered as physically a significant pressure in magnitude which corresponds to the stagnation pressure at subsonic speed with its direction coinciding with the normal to the leading edge.
3.2 Delta wing symmetrical about x axis with subsonic leading edge The velocity at a point P on a delta wing symmetrical about the x axis with subsonic leading edges (Fig. 8.3.5) is obtained by superposition of the effect of sources in the region OBPA' bounded by the leading edges OA' and OB and the Mach lines PA' and PB. The velocity induced by the sources lying in the region OAPB is determined by formula (8.3.9). The sources distributed in the region OAA' are responsible for the velocity at point P, which is calculated from the expression (8.3.13). The resultant value of the velocity is
This expression can be simplified. Taking summation and using the property of logarithm, we get
which gives
The second term under the square root sign can be expressed in the form
Thus
Expressions (8.3.18) through (8.3.20) are valid under the condition that n>l>a. The coefficient of pressure can be determined with the help of (6.1.5) for the given value of the additional velocity component:
The sources distributed on the part of the wing OUG' have an effect on the point L situated between the leading edge and the Mach wave (see Fig. 8.3.5). The velocity induced by these sources can be calculated as a sum of the velocities induced by the sources lying in the region OUG [formula (8.3.16)] and by sources distributed in the triangle OGG' [formula (8.3.1311.
Fig. 8.3.5. Delta wing, symmetrical about x axis, with subsonic leading edge.
Consequently
The first term under the square root sign may be represented by
p = 2 (u/V,), will be
300 AERODYNAMICS
3.3 Infinite semiwing with supersonic edge For this kind of wing (Fig. 8.3.6) the Mach line OK emerging from the apex lies on its surface. Consequently
(742)~c>p,, tan K < a', ?=tan ti/a'< 1 (at=cot p,=
JML1).
Let us consider the velocity at a point L on the wing between the leading edge and the Mach line OK. As the side edge of the wing, coinciding with the x axis, lies behind the Mach line drawn from point L this edge has no effect on the flow at this point. This flow will be the same as that on a flat plate interacting with the supersonic flow Vnw V sin K > a , in the direction nor= , mal to the leading edge. The additional velocity component by the expression (7.6.1) and the formula p =  2u/V, will be u=
Avw COS r
~ / . r cos2 ~ K f ~
1.
M:=af2+l we have
and 1l/cos21c=tan2rc,
from which
u =  L V,/af Jl,
where n =tan ~ / a < 1. '
Formulas (8.3.26) and (8.3.27) are applicable in cases when n< 1 and 1 >a>n. Let us find the velocity and pressure at point P lying between the Mach line OK and a side edge. If we assumed that this point lies on a wing whose apex is situated at a point C (see Fig. 8.3.6) the velocity could be obtained taking into account only the effect of the leading edge, that is, of sources distributed in the region PCH. By formula (8.3.26) this velocity becomes
To determine the actual velocity at point P on the given wing with apex at a point 0 it is necessary to calculate the velocity induced by the sources distributed over a triangle ACO from formula (8.3.28) but with the opposite sign for the intensity of vortices. The value of this actual velocity is obtained from formula (8.3.7). Replacing the upper limit of integration by zc= (XP afzp)/(tan rca') in this formula, we get
where tan rc< a' (y >,urn). Recalling that at n =tan ~ / a< 1, a =zp tan K/XP < 1 and a c n ' 4 (tan2K a'2) (x:  af2;)  4 (XPtan u a'2 zp) 2
sin'
2 (tan2 rc  af2)zc 2 (XP rc aI2ZP) tan 4 4 (XPtan rc  af2Z P ) ~4 (tan2K af2)(x; at2z ) :
 2 (xp tan rc af2zp) (xp tan u af2Z P )4 (tan2K af2)(x:  ar22 ~ ) :
1.
(8.3.30)
2 (tan2K (xp afzp) 2 (XPtan K  at2ZP)= 2af (xp zp tan K). tan Ka'
The quantity under the square root sign in equation (8.3.30) can be simplified to
Hence uAoC= af2ZP XP tan tc a' (XP ZP tan K) at2zp tan K AV, n & f ~  t a n ~ KySl xpPtan K). a' ( ~ p Z
UAOC =
Taking into account that n =tan tc/af, a=zp tan K/XP, we get
where a < n < I. The induced velocity at a point N, lying between the Mach wave OKf and the side edge, is obtained by superposition of effects of sources distributed in the region of a surface OEF bounded by the leading edge OF, side edge OE and Mach line E F emerging from point N. Formula (8.3.29) is used to calculate the velocity where the upper limit of integration is to be replaced by the quantity ZF= (xN+afzN)/(tanK +a1) and the coordinates xp, ZP are to be changed to the corresponding quantities
XN, ZN:
[
 sin'
 at24)
2 (XN tan Kat2 ZN) . 6 4 (xNtan K a12Z N ) ~ (tan2rc  a'2) (x; ar2 4 zz)
Inserting the value of ZF = (xN+afzN)/(tanIC +at2), the numerator of the first term in the square bracket will be (xN+" 'N) tan K +a'
I.
(XNtan 
zN)= 2
(zNtan K xN).
For convenience in further simplificationswe express the quantity (8.3.31) for point N in the form 4a12(zNtan K  x ~ ) ~ . Taking these values, we get
UOEF
=
"W
nl/at2tan21~ 2
[Z 
UOEF = 
A V,
n d a t 2  tan2 K
<n. where a<O, n < 1, (a1 If the positive values of ZN > O and 0 =zN tan the absolute values of n are taken,
K/XN
y cot X
Fig. 8.3.7. Pressure field for triangular panel of wing with supersonic leading edge.
u=
aa' 4 n2 1
a Vcm cos' 
n2+a n(l+c)
The pressure field for an infinite delta semiwing with a supersonic leading edge is shown in Fig. 8.3.7. The pressure is constant between the leading edge and the inner Mach line. It decreases further and reaches the value of the pressure of undisturbed flow at the external Mach line (;=O).
3.4 Delta wing symmetrical about x axis with supersonic leading edge The velocity and pressure coefficientat a point L (Fig. 8.3.8) lying between the Mach line O K and the leading edge are determined by formulas (8.3.26) and (8.3.27) respectively because the leading edge O R affects the flow only in this region. These formulas are applicable under the conditions of n < 1; l>a>n.
Fig. 8.3.8. Delta wing, symmetrical about x axis, with supersonic leading edge: IMach line: 2line
of maximum thickness.
If point P situated within the Mach angle is taken, not only the leading edge but also the side and trailing edge have an effect on the flow at that point. The velocity, governed by the effect of the leading edge and a section of the line of maximum thickness OA (see Fig. 8.3.8), is determined by formula (8.3.33). The velocity induced by the sources distributed in the region OA'A is calculated from the expression (8.3.36). Superposing all these components, we get the resultant velocity at point P on a symmetrical wing:
Considering that
4. Flowinteraction of Swept Wing with Four Corners Having Symmetrical Profile and Subsonic Edges at Zero Angle of Attack
Using the formulas for determination of velocity and pressure on the surface of a delta wing it is possible to analyze the flowinteraction of wings with symmetrical profiles and arbitrary plan at zero angle of attack. Let us take a swept wing with four corners as shown in Fig. 8.4.1. Here and in some other figures the left system of coordinates is used. for the convenience of threedimensional representation of the wing system, the sections in question and the necessary notations. We assume that for this kind of wing the leading edge, trailing edge and line of maximum thickness CBCf will be subsonic. Accordingly the angles of sweep KI and 1c3 of the leading edge and trailing edge and the angle ~2 of the line of maximum thickness will be more than (n/2) pm. The distribution of velocity and pressure over a profile depends on the position of the profile along the wing span, i.e. on the coordinate z of the wing section. Profile FL (z=zl): It is necessary to examine the four regions of flow interaction on the profile: FG, GH, HJ and JL. Region F'G is bounded by point F on the leading edge and point G lying at the intersection of a Mach line and the coordinate plane z =21. Here point G is assumed to be situated in the zOx plane and hence it is determined as the point of intersection of a Mach line from point B f , which is a projection of point B in the z O x plane,
and the line z=zl (see Fig. 8.4.1). The velocity and the pressure coefficient in region FG situated behind Mach line OKo on the wing surface are determined with the help of the distribution of sources in a triangle OCC' using formulas (8.3.20) and (8.3.21) respectively.
Fig. 8.4.1. Wing having four corners and symmetrical profile with subsonic edges.
As the inclination of a surface is equal to 1 the pressure coefficient will be 1 obtained from (8.3.21) as
1 where nl =tan ~ c l l a ' , a =zl tan KI/XI, XIis the variable coordinate of a point. The drag coefficient of an airfoil corresponding to region FG will be
CT~=
z+tan ul , x1
zl tan K I
dxl = 
zl
dx 1 0 , : tan I C ~
we have
=
 8 z, A: tan K~
b.lra'Jn21
= 1 ~
CXFG
cosh'
fi1c1
XG
.& 6'
(8.4.3)
where
g1~=
aIG=
zl tan
The distribution of sources in QOCC' with intensity of Q = 211V, has an effect in region GH and in the triangular surface BCC' where the strength of sources Q =2 (A2 11) Vm (the sign of angle 1 2 is opposite to that of ]"I). As the section G H stretches behind the Mach line OK0 within the limits of the wing the calculation of the pressure coefficient, based on the action of distributed sources in the region OCC', should be done with the help of formula (8.3.21). The effect of distribution of sources in triangle BCC' on the pressure coefficient could be obtained with the help of the relation (8.3.25) because section G H is situated behind the boundaries of triangle BCC' between Mach wave B'KB and edge BC. So
where nz=tan u2/a1, az=zl tan rtzlxz, x2 is the coordinate taken from a point B and equal to xz=xr XB. Using formulas (8.4.2) and (8.4.4) the coefficient of drag for section GH will be obtained:
CXGH = 
8 A; z, tan u,
bna' dnf  1
c1G
Ia:
a :
 8 (3211)
1 1 21 tan 1c2
.
(8.4.5)
bna'dnm where a z and 6 2 G are taken with respect to point B. ~ Summing (8.4.3) and (8.4.5) and taking into account that zt tan rcl
=X p
, Gin==21tan 1c1 XH
zl tan K I
XB 121tan ~ 2 '
308
AERODYNAMICS
21 tan ~ 2 0 2 G =tan
XG
02H =
zl tan ~2
XH
 1, =
we get
21
tan K I
='2="2
coshI
J
n:1 021
d& a :
(8.4.6)
'
Let us assume that the part of chord HL (considering that point H is located on line BC) is equal to Fb, where; is some nondimensional coefficient of proportionality determined from the condition T=B'D/~, is the root (b, chord). Then for the surface OBC the part of chord FH will be (1 7) b. As part B'D of a root chord is equal to br the remaining part OB' will be equal to (I F)b,. The angles 1 1 and 1 2 can be expressed in the following way:
where d =Alb, is the relative thickness of the profile. Taking into account the values of 1 1 and 1 2 , formula (8.4.6) could be represented in the form
CXFH =
2 b (1 F) xu'
[t a z
(17)
dnf1
02FGn2
CTIF=l
+ 
tan ~2 rdnz 1
(8.4.8)
0 2 ~ ~ 1
The velocity on a line HJ is induced by the sources of intensity Q =211V, distributed in triangle OCC' and by the sources of intensity Q = 2 (12&)V, distributed in section BCC'. The first distribution of sources establishes the pressure coefficientwhich is determined by formula (8.3.21) with the substitution of 1=11, n =nl and a =al. The pressure coefficient due to the second distribution of sources is also determined by (8.3.21). Here we must take 1 22  11, =nz and a =o2. Adding these pressure coefficients, we get = n
we get
"2J
tan 1c2
c o ~ h  l n20: do2 d
kSz
where a 2 is taken with respect to point B. For section JL the effect of a threefold distribution of sources, namely in triangular surfaces OCC', BCC' and DCC', should be taken into account. The first two distributions of sources control the pressure coefficient determined with the help of formula (8.4.9) in which and a 2 are taken with respect to the points 0 and D respectively. The additional coefficient of pressure given by the induced effects of the sources distributed in region DCC' having intensity  1 2 could be found from formula (8.3.25), in which it is necessary to assume 1 il2, n =n3, o =as. = Superposing the coefficients of pressure due to all three distributions of sources, we get
where al, a 2 and 6 3 are taken with respect to points 0, B and D. Inserting the value of & in formula (8.4.2) and replacing 1 1 by 12, we get
03J
cxHL'
2J2 2,  b X [(l
n:a: 10:
do, of
03L
n:1
do,
In this expression olH= zl tan K I zl tan K I , OIL= b,+zl tan 1 ~ 3 (1 r) b, + Z I tan KZ zl tan KZ fS2~=1,6 2 L = r b , + z ~tan ~3 zl tan I C ~ ' o3.T=1 tan K3,n3, 0 3 ~ ~ = x; XL
9
where xi=zl cot ,u=z, a', xl=zl tan K,. The integrals appearing in (8.4.8) and (8.4.13) can be solved by parts using the formula coshI a =In (a + da2 1). Then
3 11
The plus sign is taken for x = 2, n > 1 > o and the minus sign for .u =0, 1 < a < n. Let us introduce the notations:
20
coshl
J102
n2
ax
do .a2
(8.4.15)
Substituting a = 1 in expression (8.4.14) with the plus sign, we get N=ln 2. Assuming a = n in the same expression (8.4.14) with the minus sign, we get Q = (ln 2n)ln. Noting this, the resultant coefficient of profile drag is obtained as a result of summation of (8.4.8) and (8.4.13):
where Nt, N;? and N3 are obtained by formulas (8.4.14), (8.4.15) for the conditions n > 1 > o, x = 2 (plus sign) according to the values
GlH=
olL=
1 I
(8.4.17)
b 2 ~
Y)b,/(tan
rcl
(8.4.19)
or (Fig. 8.4.2) at The drag coefficient of a profile, determined by formula (8.4.16), is with respect to the local chord b. The value of drag coefficient c,, with respect to the root chord b, is obtained from the formula CXO=CX (bibr).
312 AERODYNAMICS
Profile at wing root section (z =0): The values of aerodynamic coefficients for this profile are determined in the following way: The velocity in section OB of length (1 3 b, of this airfoil (see Fig. 8.4.1) is induced by the sources of intensity Q = 211 V distributed over the triangle OCC'. According to this , the coefficient of pressure can be calculated from the formula (8.3.21). l Assuming a =z tan ICI/XI =0, we get
The sources with intensity Q = 2 1 V distributed over region OCC' and 11 , with intensity Q =2 ( 1 11)Vm distributed over a triangle BCC' act on region 12 DB of lengthb, of the profile. The coefficients of pressure are determined accordingly. Applying formula (8.3.21) with a =O, we get
The coefficient of drag of a profile, taking both upper and lower surfaces, with respect to the root chord b, will be:
3 13
we get
where cosh' n2=ln (n2+ . I n n ) . Protile FIL,: Let us find the coefficient of drag of a profile (see Fig. 8.4.1) that has the coordinate zl satisfying the inequality ZBI< ZI< ZDI. The sources of intensity Q=2A1Vm distributed over the region OCDC' and those with A distributed over BCC' act on section FlHl of this intensity Q =2 ( 2  AI) Va3 profile. Therefore it is possible to use formula (8.4.4) to calculate the pressure coefficient and the relation (8.4.5) to determine the drag coefficient. This relation (8.4.5) for the section FlHl of this profile can be written in the form
C X F ~ H= ~
cosh1
nfa: J. 1
do, of
(8.4.25)
o y
where a1 and a2 are taken with respect to 0 and B respectively. The pressure in region HlLl of the profile is determined in the same way as for region HL using formulas (8.4.9) and (8.4.11). The corresponding pressure coefficient c~.~L~ is found from the expression (8.4.13) in which the limits olH and olL are replaced by the quantities alH1 and o l ~ lthe limits , a 2 and G ~ by ~ L and ax,, and the limits osJ and o 3 by 0 3 and ~ T ~ L ~ , ~ ~ C which could be determined from formula (8.4.13'). The total drag coefficient of profile F1L1with respect to root chord b, is equal to
where Nl, N2, N3 are obtained from the expressions (8.4.14), (8.4.15) for the corresponding values of olHl atL1and 0 2 ~which are calculated for points 1
3 14
AERODYNAMICS
HI,and LIby formula (8.4.17). The quantity Q1 is determined from the relations (8.4.14) and (8.4.15) for x=O at o=z1 tan ~
2 / [ tan K I  ( l  r ) ~ 1
&I.
(8.4.27)
Profile F2L.2: Let us take the profile situated between points Dl and B1 (see Fig. 8.4.2) with the coordinate ZI which satisfies the inequality z ~ l < zl< ZB,. The flowinteraction of this profile is fairly complicated in nature. The velocity in region F2G2 is induced by sources with intensity Q=2A1V, distributed over the triangle OCC'. Hence the coefficient of pressure in this region is determined with the help of formula (8.4.1) and the corresponding drag coefficient C,F,G, is given by expression (8.4.3) in which the integration is taken in the range from a 1 to 0~ ~ ~ . ~ 1 The pressure in region G2H2 depends on the influence of sources with intensity Q=2A1V, distributed over region OCC' and also of sources with intensity Q =2 (12 A1) V , situated in region BCC'. Consequently the pressure coefficient is calculated by formula (8.4.4) and the drag coefficient C ~ G ~ H ~ is given by expression (8.4.5) in which the first integral is carried out in the range from a l ~to 0 1 ~ 2 the second integral in the range from , and a 2 ~ 2a 2 H 2 . to The sources distributed in region OCC' (Q=211V,) and those over triangle BCC' [Q =2 (A2 11) V,] and surface BCC' (Q =  2A2V,) have their effect in section H2J2 of this profile. The first distribution establishes the pressure coefficient obtained from formula (8.3.21) and the remaining two distributions give the coefficient determined by expression (8.3.25). The total value of the pressure coefficient on this section of profile will be
B where D l , o2 and o3 are taken with respect to 0, and D respectively. Applying the formula
it is possible to calculate the drag coefficient of the section of profile in question with respect to root chord b,. The flowinteraction of the last section Jz L2 of this profile is governed by the influence of the sources distributed over the same regions of the wing as those for section H2Jz of this profile. Here we must take into account the
315
fact that the velocity induced by the sources over region BCC' is determined by formula (8.3.20) where 12 is replaced by the angular coefficient (122121). Therefore the pressure coefficient should be calculated by formula (8.4.11) and the drag coefficient C X J ~ L should be obtained from expression (8.4.12) in ~ which the integrals are evaluated in the range from anJ2 cnLZ(n = 1, 2, 3). to The resultant drag coefficient is obtained by superposition of coefficients for all four sections of the profile: (8.4.30) C X F ~ G ~ CxCi2ff2 + C X H ~ J Z+ C x ~ 2 . 2 ~
where
Q2
I C ~ / tan ~  r br), ( Z ~2 provided the value of zl satisfies the inequality F b, <Z1< (1TI br tan K I a' tan ICI a1
Profile F3L3: Let us consider section F3L3 (see Fig. 8.4.1) with the coordinate zl satisfying the inequality: (8.4.32) Z D < zl< 202 ~
3
where
The pressure on section F3J3 of the profile is established by the action of ) , the sources distributed over the triangular surfaces OCC' ( Q =211 V and BCC' [Q = 2 ( 2 ill) Voo]. Therefore the coefficient of pressure can be calcu12 lated from expression (8.4.4) and the drag coefficient can be obtained from formula (8.4.5) in which the limits (TnG and s , are replaced by the quantities ~ g n and an (n= 1, 2). ~ ~ Besides the above distributions of sources OCC1and BCC' the sources of intensity Q = 2A2V, distributed over triangle DCC' also act on section J3H3 of the profile. The coefficient of pressure is equal to the value calculated from formula (8.4.4) and the additional value obtained from formula (8.3.25) in which A = A2.
J,
The flowinteraction of section H3L3 is characterized by the induced effect of the sources distributed in three regions of the wing: OCC', BCC' and DCC'. Accordingly the calculation of the pressure coefficient on this section of wing should be carried out with the help of formula (8.4.1 1). Calculating the corresponding components of the coefficient of drag for all three sections and summing them, we get the resultant coefficient of drag:
C X F ~ =~ L
7 b Y2 (1  3 Cr
It is easy to see that expression (8.4.34) is convenient for calculation of the drag coefficient of the profile situated between points Dl and D2 (see Fig. 8.4.2) and having coordinate zl which satisfies the inequality:
(8.4.35)
Profile FsL~:Let us consider section FsL5 (see Fig. 8.4.1j with the coordinates where (8.4.37) zn2= b,/(tan K I  a'). Three distributions of sources, namely OCC' ( Q =211 V,), BCC' [Q = 2 (A2ll)V,] and DCC' (Q =  212V,), act simultaneously on this section. Section F5Hs of the profile is situated behind Mach line OKo within the boundary of a wing. So we must use formula (8.3.21) with 1 = 1 1 for calculation of pressure due to the distribution of source OCC'. The second distribution of sources BCC' acts on section F5H5, which is situated between the Mach line and edge BC within the limit of the wing surface. Here it is necessary to apply formula (8.3.25), substituting 1by kz l * ~ , calculation for of additional pressure caused by the effects of distribution of sources BCC'. The second section H5L5 of the profile lies on the Mach lines OKo and B'KB, i.e. on one side of the Mach line and the corresponding edges OC and BC. Formula (8.3.21), having the distribution OCC' corresponding to the quantity 1 ,I1 and the distribution BCC' corresponding to the quantity A = A 2 1 1 , is = applied to determine the coefficient of pressare from the distributed sources OCC' and BCC'.
Section HsLs.1ies on different sides with respect to Mach line DKD and railing edge DC, i.e. behind the triangular surface DCC' where the intensity )f sources Q = 212V,. So we must apply formula (8.3.25), changing I to  A2, to calculate the coefficient of pressure due to these sources. Using the value of the coefficient of pressure obtained it is possible to letermine the corresponding coefficient of drag of profile F5L5:
cx FsLs =
zb,? (1  3 2
[NI r  r ( 1  r ) ~2;21n 2 1
dn7 1
(8.4.39)
The results of calculation of the cx/Z2 distribution of the function for drag 2.8 coefficient cx/T2 (d=d / b ) along the 2.4 span of a swept wing having cons~ tant chord b ( K I = xz = I C =60") and 2.0 symmetrical rhombus airfoil ( 112) ; = at M,= 1.8 and 1.9 are shown in Fig. 8.4.3. From the graph it is seen that as we move away from the root chord 0.8 the drag coefficient reduces. The value of the function cx/Z2 for the profile of a rectangular wing is shown 0 in the figure for comparison. To i l  0.4 Il 2 3 I4 I5 t 1 Z / b determine the resultant drag coeffi= cient of a wing it is necessary to Fig. 8.4.3. Distribution of coefficient of integrate along the span the distri drag along span of swept wing with constant chord and subsonic edges bution of drag coefficients cxk of the (dotted linesfor unswept wing). profile using the formula
Effect of side edges (tips): If the wing has a side edge (Fig. 8.4.4) it is necessary to take into account its effect on the pressure distribution and coefficient of drag. Calculation of the tlowinteraction of this kind of wing having
six corners will be done in the following way: First the velocities and pressures in region OO'D'D due to the distribution of sources in triangular section OCD are calculated. The calculation will be carried out as in the previous case (see Fig. 8.4.1), in which a wing plan with four corners without side edges was examined. But it is necessary to improve the calculated velocities and pressures to account for the effect of side edges O'C", which is equivalent to the action of sources distributed over a triangle O'CD'. The intensity of these sources will have a sign opposite to that of the sources corresponding to a wing with an area O'CD'. The section of sources distributed over a triangle O'CD' spreads over the wing within the boundaries of section OfT"D' surrounded by Mach line O'K', trailing edge and side edges. For example, for the profile F,L, the effect of sources is restricted to the region F;L, (point F; lies at the intersection of the chord FzL2 and Mach line O'K').
Let us examine how the pressure on section J2L2 of this profile is calculated. Considering only the distribution of sources in the region OCC', the coefficient of pressure can be determined by formula (8.4.1 1). The correction d p o n the influence of sources with the opposite sign in a triangle O'CD' can be carried out so that
P J ~ ~ ~ = P JAP. L
(8.4.41)
In determining the correction d p the position of section J2L2 of the profile with respect to Mach line 01'K1' passing through a point 01'which lies on the opposite side edge should be taken into account (see Fig. 8.4.4). If this line does not intersect section J2L2 the distribution of vortices in the region O'CD' of one side of the wing will affect only this section. At the same time the effect of the sources OI'C'D~'is excluded. The induced velocity is calculated by formula (8.3.13) and the corresponding additional value of the pressure coefficient d j = 2u/V,. This additional quantity can be represented in the form of the summation
where dh depends on the distribution of the sources O'CC' (Q =211 V,) and A& and A& are due to the distributions of sources B'CC" [Q = 2 (1211) V,] and D'CC" ( Q= 222 V,) respectively. Applying formula (8.3.14), we get
where a , , oz and as are taken with respect to points 0 ' , B' and Dr. If Mach line 01'Kl' intersects chord FA2 then, along with the effect of the sources O'CC", we have to take into account the effect of the sources distributed in triangle O1'C'ClU on the opposite side of the wing. The same formula (8.3.13) is used for the calculation of induced velocity.
5. Flowinteraction of Wing with Four Corners, Symmetrical Airfoil and Edges of Different Types (Subsonic and Supersonic)
5.1 Leading edge and middle edge subsonic, trailing edge supersonic Disturbances from the supersonic trailing edges of a wing (Fig. 8.5.1, a) propagate downstream within the boundaries of the Mach cone with generator DKD. As such they have no effect on the flowinteraction of a wing surface. The velocity and pressure depend on the effect of the leading edge and middle edge. Let us consider the profile FL with a coordinate z l < zgi. The coefficient of pressure at section E%, due to interaction of the sources with intensity Q= 211V, distributed over a triangle OCC', is determined by formula (8.4.1)and the corresponding drag coefficient CXFG by expression (8.4.3). The coefficient of pressure at the next section GH affected by distributed sources in the triangles OCC' (Q =2AlV,) and BCC' [Q =2 (12 A , ) V,] can be calculated by formula (8.4.4). The corresponding value of coefficient of drag C ~ G Hfor this section can be determined from expression (8.4.5). We find the interaction of the same distribution of sources on section NL as on section GH. However, considering that section HL lies on the wing surface below Mach line BKB the calculation of pressure coefficient &L should be carried out with the help of formula (8.4.9) and that of drag coefficient CXHL from expression (8.4.10). The resultant drag coefficient of this profile will be
320
AERODYNAMICS
In examining section FlLl having a coordinate zl >zsl the effect of the distribution of sources OCC' and BCC' should be taken into account for simultaneously. Drag coefficient C X F ~ H ~ section F t H l can be determined ~ L ~ by formula (8.4.25) and drag coefficient c ~ for~section HlLi can be found from expression (8.4.13) in which the third term in the square brackets is taken as zero and the limits an^ and O,L are replaced by the quantities o , ~ , and anLi (n= I , 2) respectively. The total drag coefficient of airfoil F ~ Lwill I be
Fig. 8.5.1. Wing with four corners in supersonic flow: aleading edge and side edge subsonic, trailing edge supersonic; &leading edge subsonic, side edge and trailing edge supersonic. call edges supersonic.
5.2 Leading edge subsonic, middle edge and trailing edge supersonic The main feature of the flow interaction of the wing shown in (Fig. 8.5.1, b) in that the sources distributed over the region BDDI have no effect on the distribution of velocities and pressures in the remaining part of the wing situated above Mach line BKB.
Let us take the airfoil FL with coordinates zl< ZD,. The sources distributed in triangle OCC' (see Fig. 8.5.1, a) interact on the section FH of this profile. Hence the pressure distribution can be found with the help of formula (8.4.1) and the corresponding drag coefficient CXFH can be obtained from expression (8.4.3) in which the upper limit o l should be changed to ol~. ~ The second section H G undergoes the effect of the subsonic leading edge OC (and hence of the distribution of sources OCC') and of the supersonic middle edge BC. The corresponding pressure coefficient is determined in the form of the sum of two coefficients. The first of these coefficients is calculated from expression (8.4.1) and the second from formula (8.3.27), where it is necessary to take 1=1.2121 and n=nz. Thus
The velocity on section GL is induced by sources distributed over triangles OCC' and BCC' (see Fig. 8.5.1, a). Applying formulas (8.5.3) and (8.3.40), we get the following relation for the coefficient of pressure:
The airfoil FILIis situated below Mach line OK0 and therefore the distribution of sources with intensity Q=211V, in triangle OCC' will affect it. Besides that, the distribution of sources with intensity Q = 2 (,I2 11) Veo over a triangle BCC' will act on section HILI and cause additional pressure which can be calculated by using formula (8.3.27), replacing 1 by (A2121). Accordingly, the coefficient of pressure at section F1Hl is calculated from expression (8.3.21) and the coefficientof pressure at section HILl from formula (8.5.3). The coefficient of drag of the profile will be
PF~H,
FHILl
5.3 AU edges of wing supersonic For this kind of wing (Fig. 8.5.1, c) the Mach lines OKO,BKB and DKD drawn from the points 0,B and D are located below the corresponding edges OC, BC and DC. Therefore the pressure coefficient should be calculated using formulas (8.3.27) and (8.3.40). Let us study airfoil FL with a coordinate O< z l < ZD,. The section F H lies between the leading edge 06 and the Mach line OKo. Therefore, the other edge OC' (see Fig. 8.5.1, a) will not affect the flow in this region which is considered to be a plane supersonic flow. Considering that the sources in a triangle OCC' withintensity Q =211V, interact on the section FH the pressure coefficient &H can be determined from formula (8.3.27) in which 1 and n are changed to 1 1 and nl respectively. The additional pressure on the section HG is established from the effect of the edge OC'. The pressure coefficient jHGon this section is found from expression (8.3.40) with n=nt and a = o l . Besides the interaction of distribution of sources OCC' on section GJ the distribution of sources BCC' with intensity Q =2(12 11) V gives induced , effects in this section. Applying formulas (8.3.40) and (8.3.27), we get the expression for the pressure coefficient:
The last section JL of the profile undergoes the additional effect of the opposite edge BC'. Accordingly the coefficient of pressure on section JL, where sources with intensity Q =2 (12  11) V , interact, will be
The coefficient of pressure GIHlon section FIHI of profile FILI with the coordinate ZD,< z l < ZD, is determined by formula (8.3.27) in which 1 = 1 1 and n=nl. At the neighboring section HlGl formula (8.3.40) with 1=11, n=nl and a = should be applied to determine the pressure coefficient &,G,. The pressure on the last section GtLl is obtained from interaction of the distribution of sources OCC' (Q =211V,) and BCC' [Q=2 (1211) V,]. Therefore
323
formula (8.5.7)can be used to calculate pressure coefficient &&,. The drag coefficient of profile F l L 1 will then be given by
%Fl
*X1
XGl
(8.5.10) The flow around the airfoil F&2 at a section ZD, < z l happens to be the plane supersonic flow. Therefore the pressure coefficient on it is determined by formula (8.3.27). The sources distributed over triangle OCC'with intensity Q = 2A1Vmact on the section R G 2 . Hence the coefficient of pressure ~ ' F ~ G is found from formula (8.3.27) in which 1 = d 1 and n = n l . The sources with , intensity Q = 2 ( 1 2  11) V distributed over triangle BCC', have an additional effect on section G2L2. Therefore the pressure coefficient on this section will be
From the relations obtained for a swept wing with a plan having four corners it is possible to get the relations for the aerodynamic characteristics of
a delta wing (Fig. 8.5.2). For this kind of wing the trailing edge is supersonic. The leading edge and the line of maximum thickness (middle edge) can be either subsonic or supersonic. Accordingly the local velocities and pressures and also the resultant drag coefficient cx are determined. The results of the calculation of a function cxa'/4z2 for delta wings with ; subsonic (nl > 1) and supersonic (nl < 1) leading edges with respect to (1 ) at are various values of nl (angle of sweep I C ~ ) shown in Fig. 8.5.2. The values of cxa'/4i2 at nl=0 correspond to a rectangular wing with symmetrical profile. The coefficient of drag of this kind of wing can be determined with the help of formula (7.5.22). The coefficients of pressurepl and $ entering into it , are calculated from expression (8.3.27). As the airfoil is symmetrical   pr = I pul = p and Idy/dxlr =(dyldx), = A. Therefore
where x=x/b, and the coefficient pl and p2 are determined from (8.3.27) at nl = O for 1 1 and I 2 respectively. Inserting the values of $ and $2, we get
The effect on drag of the location of the maximum thickness of an airfoil can be obtained from Fig. 8.5.2. A value of can be obtained that corresponds to the minimum drag coefficient. The points where the curves turn refer to  values of; at which the line of maximum thickness becomes sonic the (nz=nl r= 1).
The method of sources was used in Sections 3 and 5 to study flowinteraction and determine the drag force of a wing with a symmetrical airfoil at zero angle of attack, i.e. in the absence of lift force. Experiments show that the limits of application of this method in aerodynamic studies can be extended. Let us consider cases where a smallperturbation flow around a thin wing at zero angle of attack can be determined with the help of this method of sources and the lift force and drag calculated. Let us take two wings with different leading edges. One of them (Fig.
8.6.1) has a curvilinear edge with a finite supersonic region and the other wing a completely subsonic leading edge (Fig. 8.6.2).
In Fig. 8.6.1 points E,E at which the tangent to the contour coincides ' with the generators of the cones of disturbances happen to be the boundaries of a supersonic region. Consider the velocity potential at some point M situated in the region bounded by the side edge ED and the lines of disturbance emerging from points E, D,D' and B' in plane xOz. According to formula (8.2.16),in which the zone of integration o is equal to o =S1+ S2, the velocity potential at a given point will be
l which follows from In this expression the function Q (x, z)= 2 (ayl/ay),=o, (8.2.17). This function is determined from the condition of smooth flow over the wing surface (8.1.12). As the equation of this surface is given the function Q l z) will be known. In (x, the particular case of a wing in the form of a flat plate at an angle of attack a the function Q 1 = 2 V W a. So 4 determination of y from formula ' (8.6.1) is connected with determination of the unknown function Q2, which governs the intensity of distribution of sources over the region of Fig. 8.6.2. Wing with subsonic leading edge. area S2. To find this function Q2 we take an arbitrary point N(x, 0, lying in the region between the Mach lines z) drawn from points E and D. According to (8.1.20) the velocity potential at this point is equal to zero. Therefore the following expression can be obtained according to the notations indicated in Fig. 8.6.1:
"IM
The first term on the right side of this integral equation is a known function because the intensity Q on the area S 3 is determined from the boundary conditions. Therefore the unknown function Q2 representing the intensity of q sources in the region S can be obtained from this equation. So if the leading edge of a wing with a symmetrical airfoil is completely or partially supersonic the method of sources is applicable in the study of the flowinteraction of the wing at a given angle of attack. The same conclusion can be drawn for a wing with similar edges and cambered airfoil placed in a flow at an angle of attack a#O or at zero angle of attack. Now let us take a wing with subsonic leading edges. For the similar point N the following relation can be written as (see Fig. 8.6.2):
It can be seen that we get one equation with two unknown functions Q3 and Q2. By analogy with Q2 the function Q3 represents the intensity of sources on the area Ss in the region between the left leading edge and the Mach line drawn from the apex of the wing. So if the wing has a subsonic leading edge it is not possible, by the method of sources, to work out the flowinteraction of a thin wing with symmetrical profile at any angle of attack or of a wing with a subsonic edge and cambered profile at zero angle of attack or a#O. 7. Method of Dipoles The application of the method of sources to the study of the supersonic flowinteraction, as shown above, is limited to wings with completely or partially supersonic leading edges. In other cases connected with the study of the supersonic aerodynamic characteristics of wings with subsonic leading edges at a given angle of attack (or for similar wings with cambered airfoil at a=O) it is necessary to use the method of dipoles. Let us study a dipole in a supersonic flow. To do this we determine the velocity potential of the flow formed due to an elementary source and an elementary sink of the same strength Q having the coordinates x = z =C, y = E and x=c, z=(, y =  E respectively. The selected source is located above the plane y=O at a small distance E from it and the sink is situated under this
r,
plane at the small distance  E. Writing (8.2.11) in terms of finite differences, the potential due to the source and sink pair can be expressed in the form
1 + d ( x t)2at2[(y +
+(ZC)2]
021
Introducing the notation p = d ( x  t)2  a'2 [y2 + (zquantity E~ in square brackets, we get
Using series expansion of the square root and omitting the quantities of second and higher orders of smallness in these expansions, we have at2Q y ~ d a at2QysAa n {(x 5)2at2[y2+(z 5)2]}3/2 ' np3 After finding the limit of Ay, as E+O and assuming the constant value of M = Q  Eknown as the momentum (or strength) of a dipole we get the expression for a differential of the potential function of a dipole Myda d y , d i ~ = n {(x 5 )  a.2 ~ +(z  021}3/2 '
Ay, =
Integration of this over the region a, in which the effect of the dipole is felt, gives the expression for a potential function
where n is included in the function of distribution of dipoles. It can be shown that the function (8.7.1) satisfies equation (8.1.7). To do this let us differentiate (8.1.7) with respect to y:
328 AERODYNAMICS Let us now examine expression (8.7.1) for the potential of a dipole. If the quantity  112 a is included in the expression for this potential then by (8.2.12) the double integral can be considered as a potential of sources whose distribution over an area a is given by some function M. Consequently
From the comparison of (8.7.2) and (8.7.3) it can be seen that the function pdip actually satisfies equation (8.1.7) for the velocity potential. Knowing the form of a wing surface, velocity of an undisturbed flow and angle of attack it is possible to find the distribution of dipoles M and hence the potential pdip of dipoles. The derivative of the function q d i , with respect to x gives an additional axial component of the disturbed velocity u= apap/ax. Then the coefficient of pressure j= 2u/V, on the lifting surface and the lift force created by it can be culculated.
8. Flowinteraction of Delta Wing with Subsonic Leading Edge
A wing in the form of a flat plate having a triangular plan and subsonic leading edges is supposed to fall within the Mach cone (Fig. 8.8.1). The method of distributed dipoles and the corresponding relation (8.7.1) for a velocity potential of dipoles can be used to find the lift force of this kind of wing.
Fig. 8.8.1.
In Section 3 it was shown that the additional velocity u induced by the sources distributed over an inclined triangular surface with subsonic edges depends only on the function a=zaf/x [see, for example, formula (8.3.2011. This means that the velocity will be constant along the ray emerging from the point of discontinuity of a leading edge at an angle y =tan' (zlx). This ray can be treated as the generator of a cone whose apex coincides with the point of discontinuity of a leading edge.
A flow having the property of maintaining constant the velocity and hence the other parameters along the generators of this conical surface is called a conical flow. For this kind of flow in the plane y=O it can be said that additional velocity is the function of ratio zlx, i.e. u=f (zlx). From this it follows that the potential due to sources for a conical flow also depends on this ratio. Considering the flow about a triangular surface created by the dipoles as a conical flow, the potential at a point P lying in the plane y=O (see Fig. 8.8.1) can be expressed in the form where F d i P (zlx) is a function depending only on the angle of conical surface y =tan' (zlx). The potential of a conical flow at a point with the coordinates x, y, z can be written in a more general form: According to equation (8.8.1) the relation characterizing distribution of dipoles in the plane y = 0 (on the surface of a wing) is such that (8.8.3) M (59 C = 5m (A), ) where h= [It; m (h) is some function of the argument h. Let us transform equation (8.7.1). The elementary area occupied by a dipole can be expressed in the form do = d{dC=tdhd< because d l = rdh. Therefore
hg=cot
K
qdip =
IrA'cot
rn (h) dh K
r( x  s ) ~  ~ 1yL+(z
Fig. 8.8.2. Region of effect of dipoles on flowinteraction of wing with subsonic leading edges: Iregion of effect of dipoles; 2delta wing; 3parabolic curve.
where < I represents the coordinate of a dipole which may have an effect on the flow at a point P ( x , y, z ) . This dipole is situated on the curve bounding the region of influence of dipoles (Fig. 8.8.2) and is obtained as a result of the intersection of the plane y=O and the Mach cone drawn along the flow from a point P. This curve represents the parabola AlBl satisfying the equation
( ~  < 1 ) ~  a['y~ + ( z  ~ ~ ) 2 ] = 0 2 (8.8.5)
in terms of the coordinates < I , CI. The expression under the square root sign (8.7.1) can be written in the following way:
(X
(8.8.6)
where
Noting that a= 1 a f 2 , h2 >O, we find the integral on the right side of (8.8.8):
.I
.
El
1 1n[2Ja(at~+b<~+e)+2ag+b] Jdt'+b<+cJF
dg
I:'
(8.8.9)
we get
331
(8.8.13)
The difference between the logarithmic terms in (8.8.10), taking (8.8.13) into account, will be
= ln
(8.8.14)
Thus
51
(8.8.15)
) 2,dba c
.
(8.8.16)
and keeping in view that @lay= 2yar2 [see formula (8.8.7)], we can write
'7
Cot
K
Nm (h) dh.
To determine the form of function m(h) for the distribution of dipoles we use the condition of smooth flow over a wing, which states that
'1
Let us calculate the partial derivative of (8.8.20) with respect to (zlx) taking into account (8.8.21):
O=
'OIK
2 a d ~ ( v2)2 1
cot rc
lyz0
m (h) dh.
(8.8.22)
333
m (h) dh
h
It will be shown that this equation has the same form as the equation obtained by using the method of dipoles in solving the problem of the flat plate of infinite length oriented perpendicularly to the direction of an incompressible undisturbed flow. The velocity potential for the flat plate, representing part of a wing at a section AB of coordinates Z A = c, Z B = C in an incompressible flow of velocity v=V,a in a direction perpendicular to the wing (Fig. 8.8.3), can be determined with the help of equation (2.9.16), which can be rewritten in the system of coordinates shown in Fig. 8.8.3 in the form This expression determines the potential at a point P (y, z ) due to a plane point dipole of unit strength (M= 1). If the strength of the dipole differs from unity and the distribution of dipoles along the span of the flat plate is given by the function M ( y ) the velocity potential induced by dipoles on the section of the wing d y will be
Fig. 8.8.3. Incompressible twodimensionalflow past flat plate AB from delta wing in cross flow of velocity Va.
The velocity potential at a point P induced by the effect of dipoles distributed along the span of a flat plate at the section from ZA = c to z~ =c can be written in the form
The vertical component of velocity v=aqdip/ayon the surface of a wing ( y=0) will be
334
AERODYNAMICS
vY=O=
apdip
)y=o
j M
(2712
(8.8.29)
Z
C
Since the component vy=o remains constant along the span we have
Comparing equations (8.8.25) and (8.8.30), we find that they are of the same type. Therefore the possible function of distribution of dipoles m (h) in a supersonic linearized flow can be chosen so that it has the same external form as the corresponding function M (7) in an incompressible flow. To determine the form of a function M ( 7 ) we can use the solution of the problem of determination of potential function for a flat plate in an incompressible fluid flow in a perpendicular direction (see Chapter 6 Section 2). By this method the velocity potential over the plate is obtained from formula . (6.2.7). Hence the difference of potentials over the two sides of the plate will be Ap =2 V dazz2. Now we recall that an incompressible fluid flow about a flat plate is taken as the resultant of superposition of the flow due to dipoles over the undisturbed flow [see formula (6.3.6)]. Consequently the distribution of dipoles for the flat plate in an irrotational incompressible fluid flow is equivalent to the difference of potentials Ap obtained above. Keeping in view that the function m (h) for distribution of dipoles in a compressible flow has the same form as that for an incompressible flow, the expression for this function can be assumed to have the form
where H2=cot2K; L is some coefficient of proportionality. Equation (8.8.20) can be used to determine the coefficient L. Inserting (8.8.21) and (8.8.31), we get
335
To simplify the calculation of coefficient L from (8.8.32) the integration may be carried out along the axial coordinate on the wing. Keeping in mind that z =0, y =0, for the axial line we have
i"
0
dl(1af2cot2~c)sin2qdq.
(8.8.36)
'j'
0
d l  ( I  a f 2 c o t 2 ~ )s i n 2 v d P = ~ ( k )
(8.8.37)
represents the complete elliptic integral of the second kind with the parameter
The values of the integral (8.8.37) are obtained from the tables for different values of the parameter k. The distribution of dipoles can be represented by function (8.8.3) provided we replace m (h) by (8.8.31) and use the value of L obtained from (8.8.36). As a result we get
Let us find the component of an induced velocity due to the dipoles over the wing surface y=O. From (8.7.3) it follows that
i.e, the potential due to dipoles over the wing is determined by the value of the vertical component of velocity induced by the sources. Comparing (8.2.12) and (8.7.1), it can be shown that M (5, c) may be considered as a function like the function Q (5, c) determining the distribution of sources over a wing surface of given form at a given angle of attack. If the quantity n is then included in M (5, c) in (8.7.1) the following equation can be written according to (8.2.7), (8.8.39) and (8.8.40):
where h = or h =z/x. The component of induced velocity due to the dipoles will be
The corresponding value of pressure coefficient, taking into account the possible signs before the square root term, will be
where the plus sign refers to the pressure on the lower side and the minus sign to that on the upper side of the wing. The pressure field happens to be conical about the leading edge of a wing, i.e. the coefficient of pressure j =constant for all the values of z/x =constant. The lift force acting on a delta wing consists of the pressure force on the lower surface and the suction force arising from suction on the upper surface. The elementary value of a lift force acting on the area dS=O.Sxdz (see Fig. 8.8.1) will be
The total lift is obtained by integrating this over the whole wing surface
Sw x2 cot lc: =
2Y
p,
V L S , cot lc
=4rK
Cot
K
2a cot lc I p 1 dh =
fi (4
cot
'rK
K
dh dcot2rh2
'
337
Integration of this gives (8.8.44) For a conical flow the center of pressure of each of the triangular elements starting from the leading edge is situated at a distance of twothirds of the depth from the leading edge. Therefore the center of pressure of the whole wing will be located on the root chord at a point lying at a distance twothirds of a chord from the leading edge. Let us find the position of the center of pressure of a semiwing. For this we calculate the coefficient of lift of the section
br
XA
where xA=brb; b is the local chord (see Fig. 8.8.1). Using the value of p from (8.8.43) and integrating, we get where tan v =z/b, (see Fig. 8.8.1). The quantity bci is called the spanwise wing loading. The load at the root section (bci), is determined from the condition at X A =O, z=0. It is then given by (bci), =4abr cot rc/E (k). (8.8.46)
The distribution of spanwise wing loading, as seen from (8.8.45), happens to be elliptic. The position of the center of pressure of the semiwing (coordinate z,.,, see Fig. 8.8.1) is determined from the equation of moment of forces about the x axis:
whence r,, 4 br 1 cy
=
1
0
c; br dr.
Substituting the value of c; in this expression, taking the value of cy from (8.8.44) and integrating, the following value of the coordinate of the center of pressure can be obtained:
For a delta wing of lowaspect ratio ( K + Z / ~the quantity a' cot K < 1 and ) the value of the elliptical integral may be taken as E (k) = 1. Hence for this wing cy= 2an cot K. (8.8.48)
Expressing cot u in terms of the aspect ratio of the wing 1,=4 cot K, we get If the angle of sweep IC is chosen so that ~ = n / 2 p, consequently the Mach line coincides with the leading edge, cot rc=tan p, and a' cot cot p, cot ~ = 1 . In this case the elliptic integral E (k) =n/2. Therefore for a delta wing with a subsonic leading edge (8.8.49) c, =4a cot K. Since cot ~ = t a n p , = l / d ~ k 1,
we have
cY=4a / d M k  1. (8.8.50) This value of lift coefficient coincides with that of a thin profile with sharp edge in a linearized supersonic flow. It will be shown that the coefficient of lift of a delta wing with a supersonic leading edge can be described by the same expression (8.8.50). According to (8.3.40) the pressure coefficient over a wing in the region between the Mach cones will be
and on the section between the leading edge and the Mach cone [see formula (8'3.2711:
p = 1 2 a/ar61n2,
(8.8.52)
where n = tan u/al, a =z tan K/X =h tan I<. Let us determine the coefficient of pressure difference: A p =PI pu. According to (8.8.51) and (8.8.52) we get the value of this coefficient for the section between the Mach lines
 
and on the section between the leading edge and the Mach line it will be
339
1
XA
A; dr.
Inserting the value of A from (8.8.51') and integrating, we get the ? following relation for the section between the Mach lines:
  (1 +o') sin'
7t
where
0'=z tan IC/br,O < o f <
n.
Substituting the value of A from (8.8.52') and integrating, we get the ? relation for the section between the Mach lines and the leading edge
where n < o f < 1. The distribution of spanwise wing loading does not remain elliptical in this case as in the case of the wing with a subsonic leading edge. It varies linearly over the section between the Mach lines and the leading edge and it has a curvilinear variation over the region between the Mach lines (Fig. 8.8.4). The distribution of pressure in conical form is shown in the same figure. As long as the pressure field remains conical the center of pressure lies at a distance of twothirds of the root chord from the leading edge.
Fig. 8.8.4. Distribution of coefficient of pressure difference and load along span of wing with supersonic leading edge.
340
AERODYNAMICS
This integration leads to the expression cY=4a/a1, which coincides with (8.8.50).
9. Wing with Six Corners having Subsonic Leading Edges and Supersonic Trailing Edges
Let us find the aerodynamic characteristics in the more general case of a flowinteraction at a small angle of attack with a wing in the form of a flat plate with six corners in plan with subsonic leading edges and supersonic trailing edges (Fig. 8.9.1). This form of trailing edge eliminates the vortex sheet effect behind a wing during flowinteraction. The results of the solution of the problem of the flowinteraction of a delta wing with subsonic leading edges can be used to determine the velocity
Fig. 8.9.1. Wing with six corners having subsonic leading edges and supersonic trailing edges.
341
potential. Let us consider a point A with coordinates x, z situated in the region I which is'bounded by the leading edges and the Mach lines drawn from the points G and D. The velocity potential at this point can be expressed, by analogy with (8.6. l), by the following formula:
where o is the region of integration on the wing surface; the quantities Ap2 and Ap3, representing additional potentials, are determined by expressions similar to the second and third terms on the right side of (8.6.3) for the regions of integration oland cr2 (shaded regions in Fig. 8.9.1). =2AVco=2aV,, equation (8.9.1) can be rewritten Keeping in mind that Q1 in the following way:
where
which is obtained from (8.9.1). Let us introduce a characteristic system of coordinates for further transformations. In this system the axes r and s coincide with the directions of the Mach lines drawn from the leading edge of the wing (see Fig. 8.9.1): r = (M,/2ar) (x  a' z), s = (M,/2a1) (x +a' z). The characteristic coordinates of a point A (xA,zA)will be: (8.9.4)
Let us apply these characteristic coordinates r and s to equation (8.9.2). From (8.9.4) we have:
Hence ~  < = x A  x = ( u ' / M ~ [(~A+sA)(~+s)]; ) a' (Z ! =a1 (ZAZ)= (a'/Mw) [(rA sA) (r s)] . : )  
(8.9.5)
Using these expressions in (8.9.2) and remembering that the elementary area in terms of the coordinates r and s will be dcr =dr .ds sin ( 2 , ~(see Fig. 8.9.1) ~) and the limits of integration will be SB and SA, r~ and r ~ we get the equation , for a potential function:
342
AERODYNAMICS
The coordinates of sB and rc can be expressed in terms of the coordinates of points rA and sA respectively. Since the equations of leading edges in the coordinates z and x will be z= +x tan IC then, according to (8.9.411), these equations in the coordinates r and s are transformed into the form:
r = sm (for right side leading edge) Z= +X Cot K); r=s/m (for left side leading edge Z =  X cot K),
1
(8.9.8)
where m =(n  l)/(n + I), n =tan ~c/a'. Therefore the coordinates rc and s ~ can be expressed in the following way:
~ C = S C ~ = S A ~ and
S B = ~ Bm = r A m .
Replacing rc and sB in (8.9.6) by sA m and rA m respectively and carrying out integration, we get
Substitution of the values of rA and S A from (8.9.4') and the quantity gives
where h =Z A / X A or h = c/c. Comparing equation (8.9.11) with the exact relation (8.8.41) for the potential function at a given point it will be seen that for equality of results the quantity D in (8.9.11) should be equal to
Accordingly the equation that should be used to determine the potential function on the wing will have the general forin:
343
Let us usdequation (8.9.14) to determine the potential function at point A situated in region I1 which is bounded by the Mach lines from the points D and G, the side edges and part of the trailing edge. Equation (8.9.14) may be written in the following way:
By analogy with (8.9.8) Point K is situated on the side edge satisfying the equation z=E/2. In terms of coordinates r and s the equation of the side edge, in accordance with (8.9.4"), will be Then the coordinate for point K will be
YK= SK
u v m n l [ rAE (k)
SA
M1)]sArAm).
(8.9.19)
Replacement of rA,s~ and m by the corresponding values from (8.9.4') and (8.9.10) gives
Points K' and G' are situated on the side edges and satisfy the equations z= + E/2 (the plus sign for the right edge and the minus sign for the left edge). In terms of the coordinates r and s the equations of these edges, according to (8.9.4), will be r s =  M, 112 (for right edge); r s = M, 112 (for left edge). (8.9.22)
344
AERODYNAMICS
J
+M" ' ) 2 .
(8.9.23)
Integration of (8.9.21)' taking into account the limits shown in (8.9.23)' gives
,
'III=
(8.9.24)
Calculating the partial derivative with respect to XA,we get the components of perturbation velocity u= ap;/axA (n=I, 11, ZII). Then the coefficients of pressure in the respective regions on the lower and upper sides of the wing will be:
_ 2 PIT= +  . V,
8 ~ ; ~ &axA nE ( k )
+
2a'
A
( X A + z A tan
.
rc) '
(8.9.26)
I
AP/K
Here the plus sign on the right side of the equality corresponds to the lower side of a wing and the minus sign to the upper 12 side. .lo The variation in the coefficient AF/a =  p , ) / a near the side edge as a function 8 of the distance from the leading edge (as a percentage of the chord) calculated with 6 the above formulas for a wing with five corners at M,= 1.12 is shown in Fig. 4 8.9.2. The discontinuity in pressure due to the action of the corner point of a wing 2 is observed on the surface of a Mach cone  0 at the tip. The pressure remains constant Zo 4o 6 o j i = ( ~ ~ b ) ~ o Yat the section between the side edge and . the Mach cone at the tip. Fig. 8.9.2. Distribution of coefficient The coefficients of lift and moment are dp/a=(&&/a near side edge of fivecornered wing (section AA). determined from the known pressure
(5,
345
distribution:
where S, and br are the plan area and the root chord of the wing; p, and j, are the coefficients of pressure on lower and upper sides determined by formula (8.9.26). From these formulas it can be seen that the pressure coefficients jl and j will be different in sign but the same in absolute value. , =j , we get Introducing the notation =
This integration should be carried out over the complete wing plan area.
10. Wing with Six Corners having Supersonic Leading and Trailing Edges
The flowinteraction of a sixcornered wing with supersonic leading and trailing edges (Fig. 8.10.1, a) can be studied with the help of equation (8.3.1) and the velocity potential due to sources and sinks can be obtained. This equation in new coordinates r and s [see (8.9.4)] is transformed into the form
vl=
drds s I I J ( r A  r ) (sA8)
where rA,SA are coordinates of the point A in question lying on the surface of a sixcornered wing. The coordinate lines r and s are directed along the Mach lines. These Mach lines, like the lines of weak disturbances emerging from points G and D and from points G' and D' lying at the intersection of coordinate lines r and s (Mach lines) with the side edges, divide the wing surface into eight regions. The velocity potential in each of these regions is calculated with the help of equation (8.10.1). The potential function at point A with coordinates XA,ZA situated in region I (see Fig. 8.10.1, a) will be
346
AERODYNAMICS
347
From (8.9.8) it follows that the coordinate of a point C lying on the leading edge will be The variable coordinate rg' of point B' where m = (n  I)/(n II), n =tan ~cla'. situated on this edge will be The derivative of a line integration in (8.10.2), taking into account the values of the limits in (8.10.3) and (8.10.4), will be
m r s ds. ~ m ($AS)
vAm
+(sAmr,) tan'
Js~~.:.

 aV,
pi=
(XA Z A
tan K )
(8.10.5)
arJIn2
Calculating the derivative a p ; / a x Aand determining the pressure coeffi: cient j,=  2 ( a ~ I / a ~ A ) V , lower and upper sides of a wing, we find on the
$ = lf: 2 a / ( a ' . \ l l ) .
(8.10.6).
349
This expression coincides with formula (8.3.27) obtained for a wing with symmetrical airfoil'at zero angle of attack. Let us consider region 11.At a point A lying in this region (Fig. 8.10.1, b) the velocity potential is established by the sources distributed over an area OLAK bounded by the cone of disturbances with its apex at point A and the leading edges. The area OLAK can be divided into two regions OLAK" and O K K and the velocity potential at point A can be taken as the sum of the potentials due to sources distributed over these regions. Accordingly expression (8.10.1) can be written in the following form:
,
"I==[
aV,
OLAK"
The values of the limits of integration (see Fig. 8.10.1) are determined from expression (8.9.8). For points L', K and K f situated on the left and right edges respectively we have Let us carry out the line integration in (8.10.7):
where
11
OF'K
Taking into account the value of r~ from (8.10.8) and integrating, we get
d2.j.
XAa' a' 2,) m (x, +
ZA
 (x,  a'
2,)
tan'
(8.10.1 1')
'
where the plus sign refers to the lower side of the wing and the minus sign to the upper. Using the formula xY tan' x  tan' y = tan' l+xyY we have
PII"
k a'n
41 n2
(5tan'
dm),
n2+
(8.10.12')
where
This formula coincides with expression (8.3.40) for a delta wing with symmetrical airfoil and supersonic leading edges at zero angle of attack. Let us consider an arbitrary point A situated in region III (Fig. 8.10.1, c). In determining the veIocity potential in this region it is necessary to consider the sources over the surface of the wing (over the region SI=AD1D2A1 and S2= AID2 D) as well as those beyond the surface (region S 3 =A'DD"). Using (8.2.12) the expression for rp;, at a point A (x, 0 , z ) can be written in the following form:
The strength of the sources over regions S1tS2 of the wing is known and is equal to Q =2v =2aVm.Therefore
(8.10.13') where Q (5, () is the function determining the distribution of sources in the region S3 beyond the wing. If this function is known the integration can be carried out. To determine it we can use the boundary condition (8.1.20), indicating the potential function on the plane xOz in the region between the side edge and the Mach line from a point D equal to zero. For a point A' in this region the condition of the zero value of a potential function, by analogy with (8.6.2), can be written in the following way:
From comparison of this equation with (8.10.13') it follows that it is sufficient to carry out integration over region S in the basic formula (8.10.13) for 1 calculation of the velocity potential at the point A (x, 0, i.e. z),
The resultant interaction at point A in the region of sources III lying over
352
AERODYNAMICS
the region S2 of the wing surface and over the region S3 beyond the wing is equal to zero by this formula [9]. Using expression (8.10.14),transformed into the form of (8.10.1) in terms f of the coordinates r, s, we get
I
'A'
SD3
where the coordinate of the point D3 oh the rightside edge as indicated by (8.9.8). The lower limit of the integral r,, can be found from the equation of the side edge z= 112. This equation, in terms of the coordinate r and s, has the form of (8.9.17). Hence the coordinate
rlnt
'A'
+(rAs ~ mtan)
Jry~:;:
353
Noting expressions (8.9.4') for r~ and s~ and the values of m =(n  l)/(n + 1) = rn, the velocity potential
+2 (XAZAtan a'
IC)
tan'
(8.10.18)
The velocity potential at a point A in region IV (Fig. 8.10.1, d) is determined from the effect of the sources distributed over the area AJOTR. Let us divide this plane into two parts S1= OTRQ and Sz = OQAJ and determine the potential at a given point as the sum of the two potentials induced by the sources over the regions SI and Sz. Application of formula (8.10.1) gives
0
aV,
~ I V =
nMm
S1+Sz
drds
aV,
rR
dr
SA
'A
SA
ds
(8.10.20)
ST'
SJ'
The limits srl and S J , of integration are determined with the help of (8.9.8). As the points T and J' are on the right side edge and left side edge respec' tively we have
ST,=rT,/m =r/m,
s ~=r~lm rm. . =
(8.10.21)
354
AERODYNAMICS
where rx is determined by formula (8.9.23) in the form Let us carry out the integration:
 (rAsAm) tan'
2
 2aV, 
 'rA
SA
G  ( Sam) tan' ~
After substituting the values of sA and r~ from (8.9.4') and taking m =  m =(n  l)/(n I), we get
kA
~'zA)
+tan 1 J 2
. . .+
+T'(1''A)
+. . .
~2( x r + a f z ~ )
355
a'n
4 n2 1
Using the formula tan' x tan' y = tan' [(x y)/(l +xy)] for the last two terms in the above expression we get the relation
Let us now consider region V. The velocity potential at point A situated in this region (see Fig. 8.10.1, d) will be
According to this formula the velocity at point A will be induced by the sources distributed over the wing area ANPU. The inner integral in (8.10.24) with the lower limit of SAT* rm will be =
'U
Going back to the usual coordinates x and y and taking into account the values of m =(n  l)/(n + I), i = my we have
356 AERODYNAMICS
M ,
(XA
ZA tan a' (n  1)
K)
tan'
XA
at (n 1)
+T
(8.10.25')
  2 pv= +V,
.a
The velocity potential in region V I can be determined in the following way (Fig. 8.10.1, e):
pi~=s
S l
JJ +
drds 4(rAr)(sAS)
s 2
According to (8.9.4) and (8.9.7) the limits of the integrals in expression (8.10.27) will be
Integration of (8.10.27) will then give the relation for pvr in terms of the variables r ~s ,~ Knowing their relation with X A and ZA, velocity potential . this can be obtained as a function of these coordinates. Calculation of the derivative and its substitution in the formula $;,= T 2 (ap;I/ijxA)/V,
*&
357
give the followirlg expression for the coefficient of pressure in the region we are considering:
Pvr =
540
[tanl
n a' d
aV
F
SA
drds
'A
ds
'J'
S
where
The pressure coefficient corresponds to the potential (8.10.30)calculated according to the limits of integration in (8.10.31):
358
AERODYNAMICS
Pinally, let us consider region VIII (Fig. 8.10.1, g). For the points in this region the range of integration intersects with the regions 8 and 9 2 simul1 taneously: In these regions there appear disturbances arising due to flowinteraction of both side edges DE and GH. To find the velocity potential at a given point A in region VIII it is sufficient to extend the integration in the formula (8.10.1) to the region S= S1+ S2, shown shaded in Fig. 8.10.1, g. Here, the integral calculated over the region S 2 in formula (8.10.1) should be taken with the opposite sign, i s . with the plus sign. Then
rho=
Here
aV
11
S1 S1
drds sz
'A
drds d ( r A r) (sAs)
(8.10.33)
drds
dr
"A
(8.10.34)
"HI
SF 1
II
82
JJ
"1
"JrAr) (sAS )
'F;
+J J
02
;
dr
S LI
drds J J d ( r A  r ) (sAsj=
03
J=d
0
ds 4 7 s
The limits of the above integrals have the following values: rFl =rHl =sAM, 1 sLl =mr; T,
The integrals in (8.10.34) and (8.10.35) are calculated taking into account the above values of the limits of integration and then the velocity potential
359
(8.10.33) can be found. The coefficient of pressure corresponding to this potential will be:
PVIII
=a'nd~ K T n2
[talll
J In *
J(xA
(1/2z~) (n+ (8.10.37) X A + ~ ~ Z A +' (n~1) a( ~ / ) A typical distribution of the quantity A&Z =Ggu)/a, near the side edge, as a function of the distance from the leading edge (as a percentage of the chord) calculated from the formulas for a sixcornered wing (see Section 9) with straight trailing edge (i.e. for a wing with a fivecornered plan) at M,= 1.61 is shown in Fig. 8.10.2. The coefficient A& is constant in the region between the leading edge and the Mach cones starting from the apex and the wing tips over the section bounded by the side edge, the trailing edge and the Mach wave reflected from the side edge. We find a 4 discontinuity in the curve characterizing the variation in the coefficient of pressure ratio on the Mach lines. This results from the discontinuity in the contour of 2 the edge of the wing. The coefficients of lift and moment \ can be found from the distribution of pressure obtained together with formulas u I 20 40 60 80 100, % (8.9.29), (8.9.30). x = ( x / b ) 100 Effect of vortex sheet behind wing: If Fig. Distribution of the wing has a subsonic trailing edge the daa=(hp,),a side edge of near vortex sheet behind it will have an effect Gvscornered wing (section AA). on the flowinteraction of the part of the wing surface lying within the limits of the Mach lines emerging from points E and H (Fig. 8.10.3). Let us take a point M (x, 0 , z) on this part of the wing surface and study the solution of the problem of determining a potential function therein in a general form. Using the relation (8.2.16) this function can be represented in the following way:
I.
Ix
Fig. 8.10.3. Sixcornered wing with subsonic trailing edge.
Differentiate this expression with respect to XN, i.e. in a direction parallel to the direction of an undisturbed flow. The velocity potential p beyond the ' wing in plane xOz, according to the boundary condition (8.1.19), remains constant along the given direction. Hence the quantity (8.10.39), if differentiated with respect to XN, becomes zero:
Determining function Q2 from this equation and using (8.10.39), it is possible to find the potential function for the points in the section of the wing surface situated in the zone of influence of a vortex sheet IVS. A detailed study of this problem has been made by Prof. E.A. Krasilshikova in the work [9].
11. Drag of Wings with Subsonic Leading Edges
Let us consider the calculation of the drag of swept wings with subsonic leading edges in a supersonic flow at a given angle of attack. As we said
above, a disturbed flow about such wings is subsonic by nature in the direction of the normal to the leading edge. This flowinteraction is accompanied by a gas flow from the region of higher pressure to the region of low pressure (from the lower side of a wing to the upper side or vice versa) and this results in a corresponding forceinteraction on the wing. The results of analysis of the disturbed flow of an incompressible fluid around an airfoil in the form of aflat plate placed in a flow at some angle of attack (see Chapter 6 Section 3) can be used to determine this forceinteraction.
11.1 Suction force The coefficient of drag of a thin wing with subsonic leading edges in a supersonic flow is determined by the formula
is in which c , ~ the coefficientof suction force of a wing, which depends on the angle of the leading sweep x and the Mach number M,; and
CXT=
T/qmsw,
(8.1 1.2)
where T is the suction force of the finite swept wing in question; qm=p,V2/2 is the velocity head; S, is the wing plan area:
S,,,= (112) hl= (1 /4)12tan K.
The relations obtained in Chapter 7 Section 6 for a wing of finite aspect ratio can be used to determine force T. This follows from the fact that the suction force, according to (7.6.12) and (7.6.13), is determined by the variation in the axial component of velocity in the neighborhood of the leading edge of a given airfoil (wing section) and the local angle of sweep but does not depend on the condition of this velocity away from the leading edge. Using (7.6.12) we can write
i'
(8.11.3)
where coefficient c is determined from (7.6.13). Let us find, as an example, the drag of a delta wing. To do this we use expression (8.8.42) for a perturbation velocity. Assuming the value of cot x = Z/XL.E this expression, we can write in
where XL.E is the distance from the leading edge; z, x are coordinates of a point on the wing (Fig. 8.1 1.1). Inserting the value of u2 in (7.6.13) and recalling that lim x/x~.= I, we get =
W L .E
n p, cot 2
112
2dl+tan2.ML
Jzdz o
E (k)
41+tan2 ICML.
Fig. 8.11.1. Suction force of delta wing with subsonic leading edges.
For a supersonic leading edge (M, cos K > 1) and also in a case where the leading edge is sonic (tan K =cot pm = dMTO3the coefficient of a suc1) tion force is equal to zero. But in the case of a subsonic edge (M, cos K < 1) the quantity c ~ T # O : its actual value, shown by experiment, is less than the calculated one. This is quite significant at high angles of attack or large sweep, at which local separation takes place around the leading edge and no further growth of suction is observed.
11.2 Induced drag Let us consider the total drag of a wing. According to (8.11 .I), (8.1.4') and (8.8.44) this can be represented in the following form:
Cx =
1

Let us consider the case of a sonic velocity of flowinteraction. If M,+ 1, E (k)+ 1 and hence the drag coefficient is It follows from this expression that the drag of the wing decreases with an increase in A, while maintaining the condition of subsonic flowinteraction of the leading edge. In the limit where ,I,+ca the drag tends to zero. This brings us to the wellknown Eulerd'Alembert paradox to the effect that there is no drag for an infiniteaspect ratio wing in an ideal gas flow. From this it follows that the appearance of drag depends on the wing having a finite span. Expression (8.11.7) coincides with formula (6.4.19) for the coefficient of induced vortex drag of a finitespan wing (see Chapter 6 Section 4). So the physical nature of the drag force appearing during the flowinteraction of a delta wing with subsonic leading edges is based on the induced effects of vortices felt behind this wing. Accordingly the quantity determined by (8.1 1.6) is known as the coefficient of induced drag. If the leading edge is sonic (tan K = ~ M L I) the second term in the square brackets of (8.11.6) is equal to unity and E (k) = E (0) =n/2. Hence the coefficient of induced drag will be For a very narrow wing (rc+n/2) we may assume cot 1cw0 and E (k)w E (1) = 1. In this case expression (8.1 1.6) coincides with (8.1 1.7).
The disturbed flow on the wing in region I, situated between the leading ' edge and the Mach cones, is set up solely by the leading edge. Here the pres ' f sure coefficient is determined by the formula for a flat plate
The method of sources can be used to analyze the disturbed flow in region II (see Fig. 8.12.1, a). The velocity at a point A lying in this region is induced by the sources distributed over wing section ACOE. Meantime the resultant influence of sources BCO and OBD at point A is equal to zero and hence it is necessary to carry out integration in (8.10.1) over the region ABDE. Accordingly the velocity potential at point A will be
aV, V I I= nM,
,
 
ABDE
fl
.+
I d G Id G '
B .
ds 
dr
(8.12.2)
; F
Point F lies on the leading edge, satisfying the equation x =0 Therefore . according to (8.9.4") we have for this edge
Hence the equation of the leading edge in terms of the coordinates r, swill be
Therefore the coordinate of point F is Point B lies on the side edge and satisfies the equation z=O. Therefore according to (8.9.4"), for this edge Hence the equation of side edge in terms of the coordinates r, s will be
The coordinate of point B will be Carrying out integration in (8.12.2), taking into account the values of the limits r p = s and SB =r ~ we get ,
a %,I ax,:
>
On the boundary of two regions I and ZZ separated by the Mach line the coefficient of pressure determined by formula (8.12.8) must be equal to the value of (8.12.1). In practice, since the equation of this boundary
ZA =X A
tan p =x~la', ,
It follows from (8.12.8) that at the wing tips the pressure along the line z/x=const remains unchanged. These lines can be taken as the generators of the conical surfaces and hence the flow on this region of the wing is treated as a conical flow. Let us calculate the aerodynamic coefficients of the wing. The lift force, acting on an element dS of the end section of the wing, is equal to
d Y= (P, VL/2) dS. It can be seen from Fig. 8.12.1, a that
(51 a
(8.12.9)
Inserting the values of pl and $ from (8.12.8) and the quantity dS from (8.12.10) in (8.12.9) and replacing zA/xAby tan y, we will have a' tan y ..dy : d Y 4ab2 . P, v tan* = nu' 2 1 a' tan y cos2 y Dividing this expression by the product of the wing area SIz= b2/2a' enclosed and carrying out the in the Mach cone and the velocity head q,=p,VL/2 integration, we get the coefficient of lift of the wing tips:
&W
Y c==8UI
n
SII qW
.  , dy
cos2 y
(8.12.1 1 )
Let us find the lift coefficient cYzrof the endsections of the wing with respect to the total wing area Sw=lb:
Cyrz=Cy
, ,
yka"
1 
where I , = l/b. The coefficient of lift of section I of the wing taken with respect to the area SI of this section will be c;
= 4a/a1 4 a l d M ; =
 1.
(8.12.12)
Then the coefficient of lift of this section with respect to the total wing surface area is
4a 1 d M m ('2I,, J M m ) '
(8.12.14)
The coefficient of moment about the z axis can be calculated from the available data on the distribution of pressure and the lift coefficients of various sections of the wing:
mz=Mz/qmS,, b.
(8.12.16)
Moment M, due to the pressure forces distributed over the wing surface can be found as the sum of moments of the lift forces about the z axis that are acting on the different sections of the wing (Fig. 8.12.2). Since the point of exertion of lift force on the triangular sections of the wing coincides with the center of gravity of the area of the section, the value of M, will be (sign of the moment is taken as positive for rotation of the wing from the x axis to the y axis): . b b 26 Mz= 32Y$" + Yi2'+j 2Yrr, (8.12.17) 2 yj2) where Yjl), and Y,, are values of lift forces calculated for the areas of sections CDB (HFE), E X B and ADB (GHE) respectively. According to (8.12.11) and (8.12.12):
Inserting (8.12.17) and (8.12.18) in (8.12.16) and keeping in mind the value of 5'11/5', = 1/(2 h,a'), we get
If the number M, of the flow over the wing shown in Fig. 8.12.1,a is reduced then at some value of this Mach number the tip Mach cones intersect on the wing. There then arises region III (see Fig. 8.12.1, b) in which the leading supersonic edge and both the side subsonic edges have an effect on the disturbed flow. The nature of the flow in regions I and II (IT) will be the same as in the corresponding regions I and II (TI')of the wing whose scheme is shown in Fig. 8.12.1, a.
Let us consider the flow at a point A in region 11 1 (see Fig. 8.12.1, b). The ; region of effect of the sources on this flow coincides with the section of wing ABDD'B'. This section can be represented as the sum of areas HBDD' and j ' AHD'B'. Then the velocity potential at point A, according to (8.10.1), will be
 aV,
HBDD'
drds
AHD'B'
The limits of integration can be determined directly from Fig. 8.12.1, b and the above potential can be written as:
The first two integrals are calculated in the same way as (8.12.7); the remaining two integrals are determined independent of each other because the region of integration represents a parallelogram (see Fig. 8.12.1, b). Taking this into account, we get
+ J2rA (SA r
~  (sA+rA) tan)
drA+SH
SASH
It can be seen from Fig. 8.12.1, b that the coordinate sH= sDtand accord= = = ing to (8.12.4) the value of sDf rD'. Therefore sH  rD1 rB'. Point B' lies on the right side edge, satisfying the equation z = l . From (8.9.4") it is possible to find the equation of this edge in terms of the coordinates r, s: rs= M,I. Then for point B' r ~=SB' Mml=sA MwI. ' Accordingly the coordinate (8.12.23) (8.12.22)
sH= rBf= sA+M mI. (8.12.24) Substituting the coordinate r and sA from (8.9.4') and rBf and SH from A (8.12.23) and (8.12.24) in (8.12.21), we get
369
Applying the formula = T (2/V,) ap'II,/ax~, find the coefficient of we pressure in the form
>
Let us express the second term in the square brackets in the form tan' Then ( x A + ~ ' z A )  a ' l=n 2 tan'
XA
(XA+Q'
2a (1  Z A ) a' ZA)a' 1 F ay
The first and second terms in this expression represent the coefficients of pressure > I I and for the end sections of the wing respectively. The third term indicates the pressure coefficient FI in region I where the effect of the side edges is not feIt. So the pressure coefficient in region III can be expressed in the form
P I I I = P I I +PII'
6.
(8.12.28)
Expression (8.12.28) can be obtained from the following physical considerations: The coefficient of pressure at point A in region III changes under the influence of the left side edge with respect to coefficients & by the quantity d s l = p l  k I . Similarly, under the influence of the right side edge it changes by the quantity dFII'=&&~'. As a consequence of the simultaneous effects of both side edges the coefficient of pressure at a given point changes by the quantity d ? I I r =d j I I+ A&I'. Substituting the values of d p f I and A&' obtained and using the relation A&II=PIPIII, we arrive at relation (8.12.28). The coefficientsof lift and moment are worked out from the known distribution of pressure with the help of formulas (8.9.29) and (8.9.30). The coefficient of drag of the wing in question is then worked out from the expression cn= ac,..
method of reversal of flows. This method establishes the relations between the aerodynamic characteristics of two thin wings of the same pian placed in I % opposite postures. c It is assumed that one of these wings is placed in a flow to the right at an , angle of attack at and the other in a flow in the opposite direction at an angle a2 (Fig. 8.13.1). According to the theory of linearized flowinteraction the p,V,u and the pressure difference at any point on the surface ispp,= resultant pressure between the lower and upper surfaces is
According to this expression the total drag force for wing 1 placed in a flow to the right will be
The values of the pressure difference Apl, Ap2 and the local angles of attack a , , a2 are measured at one and the same point. We will not consider suction force, which may arise at the subsonic leading edge but does not alter the results of the above calculations. As the flows around these wings are smallperturbation flows the superposition of one upon the other yields some new flow with parameters satisfying the linearized equation for velocity potential. Let us carry out superposition of the flows in such a way that the undisturbed flows cancel each other. Then the new wing 3 (see Fig. 8.13.1) will coincide with the picture of wings I and 2 with superposed flows but the camber of the surface will be different.
37 1
This camber is determined by local angle of attack a3 which is equal to the sum of the angles of attack:
a3 = v3/Vw=(v~/Voo) +(vz/VW)=a1 f az. (8.13.4)
According to (8.13.4) the vertical components at the corresponding points of the surface are added. The horizontal components of velocity will be cancelled out by superposition so that for wing 3 we get Then the differential pressure between the lower and upper sides of this wing will be
Ap3 = pwVmA~3 pwVw ( A u ~ Auz) = 
or according to (8.13.1)
Ap3 =API  Apz. (8.13.5)
The existence of a drag force for the wing in question is related to the appearance of a perturbation flow behind it representing an infinitely long vortex sheet moving away from the trailing edge. From the physical considerations it must be clear that there is a close relation between the disturbed condition of flow behind the wing and the drag force. Specifically force XI and the disturbed flow condition behind wing 1 and force XZ and the disturbed flow condition behind wing 2 stand in a direct relation. If we take wing 3 the flow past it obtained as a result of superposition of the flows past wings I and 2 will have the property that ahead of it the disturbances will be the same as for wing 2 and behind it the same as for wing I. In other words the drag force of wing 3 is equal to the difference between the drag forces of wings I and 2, i.e.
This relation represents the basic equation of the method of reversal of flows. Coming to the coefficients of pressure in (8.13.8), we can write this relation in the form
 
Using the right of arbitrary choice of angles of attack we take the values a1 and a2 as constant for all points of the surface of the wings and assume that they are identical, i.e. we will consider that a1 =a2=a. Then from (8.13.8') we find that
So the method of reyersal of flows says that the lift force of a flat plate in a flow to the right (YI) and the reverse flow (Yz) will be identical during flowinteraction if the angles of attack and the velocities of undisturbed flow are the same. Experiments show that the basic relation (8.13.8) of the method of reversal can be used for study of the aerodynamic characteristics not only of isolated wings but also of flight vehicles which in themselves represent "thin" combinations of wing with other structural elements like fuselage (body of revolution), tail and control surfaces (see Chapter 11).
The problem of the flowinteraction of a sharpnosed cone is very important for aerodynamics. The solution is of great practical value in that it helps in working out the aerodynamic characteristics of flight vehicles and their elements of conical shape. Besides this it can be used for study of a supersonic flow past sharpnosed rotating bodies. For example, it gives the initial point on a curve for distribution of the parameters of the flowinteraction of a sharpnosed curvilinear body. In addition, the results of the symmetrical flowinteraction of cones are applied in the approximate calculation of the distribution of flow parameters over the peripheral surface of rotating bodies (method of "local cones"). The same results are used for comparison in the investigation of the aerodynamic properties of bluntnosed cones. Along with the exact solutions a series of approximate solutions facilitating a simplified analysis of the flowinteraction of a cone is worked out in theoretical aerodynamics. Some of these solutions deal with slender cones interacting with a linearized flow or with a flow at a very large Mach number M. The exact solution may, in general, be applicable to cones of arbitrary thickness and the flows interacting with them may have any velocity. The basic condition to be satisfied is connected with observing the conital flow past the body in the flow. This kind of flow has parameters that remain constant along the lines emerging from the apex of the body in an inviscid flow. However, the results so obtained are also applied in studying viscous flowinteractions. "Inviscid" parameters like pressure, velocity, density are regarded as the parameters at the external limit of the boundary layer formed over the cone and hence they are the factors that determine frictional stress and heat flows from the gas to the wall. Let us take a cone with semiangle BK at its apex in an axisymmetrical supersonic flow. The main problem is to work out the gas flow between this
373
374
AERODYNAMICS
cone and the shock appearing in front of it, which also takes the form of a conical surface. Here it is necessary to determine the angle of inclination Oc 6 of the linear generator of a conical shock (Fig. 9.1.1). For this purpose we examine the system of equations in the spherical coordinates 1 and r 1 3 applicable to the case of flowinteraction where the gas behind a shock undergoes physicochemical changes under the effect of high temperatures. We will assume here that thermodynamic equilibrium is established in the disturbed region of flow.
Fig. 9.1.1. Velocity components on conical surface: Icompression shock; 2cone in flow: 3conical surface.
The solution to be obtained for a cone must correspond to the axisyrnmetrical conical field of a disturbed flow in which the gas parameters are constant along the lines emerging from the nose that constitute the generators of the interspace conical surfaces including the conical surfaces with angles 8=8, and 8 = / 3 ~ . On the basis of this property of the solution sought, any partial derivative of the gas parameters with respect to the spherical coordinate r (see Fig. 9.1.l) is equal to zero. Then the equation of continuity (2.4.37), in which the partial derivatives with respect to the spherical coordinate rand also those with respect to y (twodimensional flow) should be equal to zero, will have the form dp dVo 2pV,+Vo +p+pVe d O dB cot 8=0.
375
The equations of axisymmetrical motion about a cone obtained from the system (3.1.45), with viscous terms and partial derivatives with respect to t and r assumed to be zero, are written: dV,/d@=Ve; (9.1.2)
According to the number of parameters to be obtained it is necessary to add the following equation to these equations:
which is obtained from the equation of state (1.4.8) for a gas at an arbitrary point in the flow and from the equation of state pc= pcTcR~/(pav), related to the conditions immediately behind the shock wave (index "c"). The energy equation (3.4.14) and the general relations of the type (4.2.8) through (4.2.1 1) for the calculation of enthalpy, entropy, average molecular weight and sound velocity (9.1.5) i (V2/2) =ic (Vz/2);
must be included in the system in question. In this form the system can be used to study the flowinteraction of a cone in a dissociating gas. The system is simplified in the absence of dissociation. If it is assumed that the specific heats and average molecular weight of a gas in a disturbed region between the shock and the surface of the cone remain the same as those in an undisturbed part of the flow, the sound velocity and enthalpy will then depend only on temperature. Then instead of equations (9.1.6) through (9.1.9) we will have to use the relations (4.3.1) through (4.3.4). These can be represented in the following form:
2. PlowInteraction of Cone at constant Specific Heats The results obtained for the flowinteraction of a cone at constant heat coefficients have great practical value. They can be used for an approximate evaluation of some flow parameters (for example, pressure) when the flowinteraction is accompanied by a considerable amount of heat productive of physicochemical changes and consequent changes in specific heats. Equations (9.1 .I) through (9.1.3) and (9.1.10) through (9.1.13) can be used for the solution of this problem. Equation (9.1.3) will come in handy if some changes are made in it. For this purpose we use the expression of the speed of sound a2=dp/dp, written in the form
Inserting here the value of the derivative dp/d6 obtained from the equation (9.1.1) and rearranging terms, we get
'
Examining the system of equations (9.1.2), (9.2.2), (9.2.3), we find that the problem of the flowinteraction of a cone is reduced to the kinematic problem connected with determination of the velocity field in a disturbed flow around a cone, i.e. with the determination of functions V,(B) and Ve(0) as the components of velocity or function V(B) =2/ V + v; for the total velocity. : From the calculated total velocity it is possible to determine pressure, density, temperature, enthalpy and entropy of the gas with the help of (9.1.4), (9.1.5), (9.1.10), (9.1.1 1). The relations (3.6.26), (3.6.31), (3.6.33) can be used to find pressure, density and temperature respectively in place of the above formulas. The boundary conditions on the basis of which we integrate differential equations (9.1.2.) and (9.2.2) are determined from the conditions of gas flow past a cone and from the conditions characterizing the gas parameters immediately behind the shock. The boundary condition of the flowinteraction of a cone emerges from the fact that the normal component of velocity over its surface is equal to .  zero, i.e.
377
We have two conditions for shock. The first is obtained from the equality of tangential components ahead of and behind the shock, i.e. Vcroa=Vcr (Fig. 9.2.1). Accordingly Vcr = V cos Bc. m (9.2.5)
Using this expression it is possible to get the second condition. For this purpose we write the expression for the horizontal component of velocity Uc of a gas at the shock (see Fig. 9.2.1): Uc=Vcr cos 0,  V,, sin 6,. Multiplying both parts of this expression by VCrand recalling (9.2.5), we get (9.2.6) UcVW vcr(Vcr V,, tan Oc), = where Vcr and Vco are the tangential and normal components of velocity at the shock respectively.
Fig. 9.2.1. System of velocity triangles in front of and immediately behind compression shock for cone in supersonic flow.
Now we use equation (4.4.4) for a shock polar and write it in the form
where Wcis the vertical component of velocity on the shock. Taking into account (4.4.3), this can also be written as
Recalling that tan2 8,=c0s~ 0,  1 and that the quantity V,Uc is determined from (9.2.6), we find
1 2  = .
COS~
ec
v2
Vzr  , V tan 0 IV o
k+ 1
378
AERODYNAMICS
'Keeping in mind (9.2.5), we finally get the boundary condition at the shock:
Integration of the system of equations (9.1.2), (9.2.2) and (9.2.3) can be carried out by a numerical method. Here the angle of shock Oc and velocity of undisturbed flow V are taken as known. The velocity field is determined , in the course of solution of these tquations and the corresponding cone angle /IKand velocity V,= VK are found. Let us see how a problem is solved. The radial component of velocity can be determined from (9.2.5) using given values of 8, and V : , (9.2.8) Vcroo Vcr = Voocos Oc. = The velocity is identical for the conditions ahead of the shock and immediately behind it. It is denoted by Vrl= V cos O,, where O1=Oc. The , normal component of velocity Vce=Vel behind the shock is obtained from (9.2.7) on the basis of this value of Vc,= Vrl:
Let us take an intermediate conical surface, near the shock, with inclination of the generator 8, A&, where A01 is a small rise in angle 8 (Fig. 9.2.2).
The radial component of velocity Vr2 past this surface can be calculated from equation (9.1.2), written in terms of finite differences: VrVr1 =VelAO. (9.2.10) Assuming here that V, = Vr2 and de =A81 =81 82 ,we get Vr2=VrlfVe1 AO1. (9.2.10')
379
The normal component VOZ be obtained from equation (9.2.2) written can in terms of finite differences:
where the derivative (dVeld9)l is calculated from equation (9.2.2) using the parameters on the shock:
1
?)',
(9.2.12)
Assuming the values obtained for Vr2, Ve;?and also a;, which is determined by (9.2.3) in the form
as the initial data, it is possible to find Vr3, Ve3, a3 on the next intermediate surface with an angle of inclination of generator
where the derivative (dVe/dB)ll,l is obtained from (9.2.2) using the parameters Vetn17 Vr,f, nn,_l and B1n1=B1,z2dBtn2. The calculations are complete when, for some value of angle of intermediate cone (see Fig. 9.2.2), i=n1 (9.2.17) 8 n = f l n  1  ASl1 ~ 0 1 d0i)
C
1
the normal component of velocity Ven does not become zero, i.e.
Here the derivative (dVe/dB)nl is found from (9.2.2) using the values of Venl, V,,I, anI on the neighboring. intermediate conical surface with an angle of inclination of generator
i=n2
In calculating it is not usually possible to select such a small angle dBn1 as a first approximation that the equality Ven=O is satisfied. Usually for the selected value of there exists a calculated value of Yen which changes its sign to the opposite of that of Ven on the surrounding surface with an angle of inclination of generator This indicates that the value Ven=O corresponds to an increase in angle less than the one selected. To determine this increase in angle it is necessary to carry out the operation of interpolation using the equality
The calclllation of velocity and angle on the cone are based on this value of
dB,1:
(9.2.20)
Similar calculations can be carried out in reverse order, imposing conditions on the cone. In this case we have to know the angle PK and the velocity V on the cone. Numerical integration is supposed to be complete K when the boundary conditions (9.2.5) and (9.2.7) are satisfied. On the basis of the gas parameters in the perturbed region the angle of inclination of shock ahead of the given cone and the velocity (Mach number) of undisturbed flow are found. Each operation of the numerical integration for given values of Bc, V (or , PK, VK)provides a possibility of determining the velocity field, i.e. of finding : the nature of function V= V(0) ,where V= 1/V,2+ V, and of establishing the relation between the given angle of cone PK and the velocity on it VK, or between the angle of shock 6, and the velocity V from the other point of , view. Repeating the calculations for different given angles Oc and a h e d value
38 1
of velocity V it is possible to find relations of the type p ~ = f i ~ ( O c ) , , VK= VK (/IK) or VK= VK (&). The results so obtained can be represented graphically in a hodograph plane W, U in the form of a socalled "appleshaped" curve (Fig. 9.2.3, a). This curve is the geometrical location of the ends of velocity vectors 7~ a perturbation flow right on the cone in a flow. for
Point A situated on the "appleshaped" curve and lying at the end of a velocity vector, corresponds to the cone with a given semiangle fig. The point B located on the shock polar coincides with the end of velocity E, on the corresponding shock with angle of inclination 0,. The curve AB happens to be the hodograph of velocity, i.e. it is the geometrical location of the ends of velocity vectors in the region of disturbed flow between the cone and the shock. A hodograph should be developed in the following way: The velocity components for the intermediate conical surface can be determined from Fig. 9.1.1 with the help of a velocity triangle behind the shock: V,=V cos (88); (9.2.22) Ye=V sin (68).
In the process of numerical integration using values of V , and Ve obtained for given angles the ratio (9.2.23) Vr/Ve= cot (8 E ) , is determined, from which the angle of inclination E to the cone axis of a velocity vector can be worked out. The polar coordinates V and c determine the positions of points on the hodograph of velocity (see point Cin Fig. 9.2.3, a). The graphic method of solving the problem of supersonic flow over a circular cone is developed by Prof. A. Bussman. The physical picture of a supersonic flow of gas over a cone can be clearly presented with the help of the appletype curve and the family of hodographs of velocity. A phased isentropic compression of gas takes place in the region between the shock and the cone along the stream line. In Fig. 9.2.3, a the interchange from point B on a shock polar along the hodograph of velocity to point A on the appletype curve corresponds to this. As can be seen from Fig. 9.2.2, the stream line gradually bends and approaches the cone surface following the direction of the generator of a cone. Let us draw an arc of radius a* with point 0 as the center (see Fig. 9.2.3, b). If the hodograph of velocity AB for a given angle of a cone lies on the right side of the arc the isentropic compression behind the shock wave the part of takes place at supersonic speeds. For the angle of cone &2 > the hodograph GK can be on the left side of the arc and the part KD on the right side. The disturbed flow will be a mixed one. The flow will be supersonic in the region around the shock and subsonic near the surface of the cone. For a still higher angle of cone BK3>BK2the hodograph of velocity EF lies to the left of the arc a* and hence the disturbed flow will be completely subsonic. Analysis of the appletype curve shows that for each angle of cone / 3 theoretically there exist two solutions (see points of intersection A and A' ~ of appletype curve A 0 in Fig. 9.2.3, a). One solution gives a low velocity and a high angle of shock inclination and the other a high velocity and a low angle of shock inclination. As shown by experimental research, the second solution, representing steady flow behind the shock, corresponds to reality. It can be shown that the ray from the origin of coordinates passing through point TA is tangential to the appletype curve. This point TA theoretically corresponds to the unique solution and determines the critical angle of cone f i ~ . c r . If the actual angle of the cone is more than the critical onerit is not possible to investigate the flowinteraction of the cone formally with the help of this curve. Under real conditions this flowinteraction is characterized by the fact that the shock becomes detached from the sharp nose and gets bent. This kind of flow over a cone is called supercritical.
It can easily be seen that the critical angle happens to be the function only of undisturbed velocity (corresponding to Mach number M,= V,/a, or the relative velocity I, = V,/a*). When the cone angle is near the critical angle (PK< j k c r ) the nature of the flow, as shown by experiment, diverges from the theoretical one. According to the appletype curve the attached straight shock and conical field of velocity (pure subsonic or mixed) corresponds to the socalled transcritical regime of flowinteraction (points E, G in Fig. 9.2.3, b). However, observation of flows over cones with cylindrical tail sections in wind tunnels shows that in reality the shock remains straight near the nose and gets bent only at some distance from it. For the same cone the point where the shock begins to bend moves nearer the nose with a reduction in Mach number M,. This bending of the shock indicates that the flow does not remain conical. This is well seen in Fig. 9.2.5, a, Fig. 9.2.4. Cone in supersonic where the flow pattern for a conical body flow. with angle P ~ = 2 5 'at Moo=1.4 is shown. The bending of a "sonic" line indicates divergence from a conical flow. It may be noted that near a sharp nose this line, though it does not fully coincide with the theoretical line obtained from the hodograph of velocity, remains straight, which indicates that the conical nature of the flow is maintained. At some distance from the nose the "sonic" line gets bent, indicating that the conical velocity distri'oution is disturbed. With a reduction in Mach number M, to 1.27 the subsonic region increases and the "sonic" line, as seen from Fig. 9.2.5, b, moving away from the apex of the cone, approaches the shock, which is already somewhat bent near the nose. A conical flow is no longer possible all over the space between the shock and the body. With further reduction in M, we.get the conditions for complete detachment of the bent shock from the nose of the cone. Experimental data show that the shock moves away from the apex at angles PK somewhat greater than those determined by accurate theoretical flowinteraction calculations for a cone. For example, for Mach number M, =2.45 it is found by experiment that detachment of shock takes place at angle &=46", whereas the calculated value of this angle in theory is 45" 46'. Experimental research shows that a conical flow corresponding to constant velocity past the cone is observed until the velocity over its surface reaches the speed of sound. The corresponding angle of cone PL can be determined by the point A" of the intersection of the arc of radius a* and the appletype curve (see Fig. 9.2.3, b). So in practical cases it is possible to use the appletype curve for calculation of the flowinteraction of a complete
384
AERODYNAMICS
As far as the apex of the conical surface is concerned the velocity past it obtained by conical theory (with the help of the appletype curve), tallies well with the experimental data for all values of angles /?KG PK.cr, i.e. for those conditions of flowinteraction under which the shock remains attached. The appletype curve gives the relation for velocity VK past a cone and angle of inclination of shock 0, as a function of cone angle Pg at a given velocity of undisturbed flow V, (at given Mach numbers M, or A , To get ) . the appropriate relation at other Fig. 9.2.5. Position of compression shock and values of velocity V , (M, or "sonic" line near cone attached to cylinder: it is necessary to carry out the aat M,= 1.4, 6at M,= 1.27; Ilinear numerical calculation of the flowshock (theory); 2curved shock (experiment); interaction of the cone and con3'csonic" line (experiment); 4'ssonic line (theory). struct the corresponding appletype curve for the new conditions. The family of such curves represents the relation of velocity on a cone and the angle of inclination of shock as functions of cone angle PK and velocity V , (M, or A , Fig. 9.2.6 shows a family of appletype curves drawn for ) . The corresponding different values of the relative velocity I,=V,/ah. shock polars are shown in the same figure to facilitate comparison of the flowinteraction of a wedge with bodies of similar semiangle PK,e.g. a cone. It can be seen from Fig. 9.2.6 that the velocity past a cone is higher than that past a wedge (07~ For a wedge the flow turns through an angle > oxw). Pc=Pw at an oblique shock. At the same time the flow is turning by an angle less than the cone angle, P, < PK. Consequently the angle of inclination of the main shock wave ahead of the wedge Oc.wis more than the angle O,.Kahead .~~ of the cone. Further , 5 ' ~>Pw.cr. The same result can be obtained directly from Fig. 9.2.3, a. Let us join points A and B with a point 0 2 and draw perpendiculars from this point to w the straight lines so obtained. Then the angles between the perpendiculars and horizontal axis U determine the inclination of the generators of shocks It ahead of wedge B,., and ahead of cone t?,.~. can be directly seen from the figure that > O,.K. The important factor of calculation of the flowinteraction of a cone happens to be the determination of pressure, density and temperature from
< 8;.
the values of VK and 0, obtained. Keeping in view the isentropic character of the gas flow behind the shock, the corresponding relations (3.6.26) through (3.6.33) can be used. Assuming that the temperature TOin formula (3.6.33) does not change behind the shock and is determined by (3.6.35), we get the following formula for temperature on a cone:
= where FK VK/Vmax. The stagnation pressure in formula (3.6.26) should be calculated taking into account the loss across the shock. Denoting this pressure by pb and ~, introducing the parameter ~ = p b / calculated on the basis of stagnation pressure po before the shock from formula (3.6.29), we get the pressure over a cone as
The parameter vo =p;/po depends on Mach number M, and angle of inclination of shock & ahead of the cone and is determined with the help of (4.3.21) and (4.3.22) in the following form:
The density ratio is obtained from the equation of state PK/P~=(PK/PCO) (T,/TK). The pressure coefficient on a cone will be (9.2.27)
(9.2.28)
The drag force based on the action of pressure (wave drag) is determined with the help of the relation (1.3.2). Taking p =p,, cfx=O, S = S, = nR$, , dS=2nRdl, cos (Fnx)= sin have
A
PK (all
Recalling that dl sin PK=dR, we get the following expression for drag coefficient:
where R =R/RK. As the pressure coefficient jK a cone in a supersonic flow is a conon stant quantity we get the following relation for the wave drag coefficient of a cone:

CxW=P K.
(9.2.30)
So the coefficient of wave drag during a supersonic axisymmetrical flowinteraction is equal to the pressure coefficient on its surface.
3. Effect of Balanced Dissociation and Ionization of Gas on FlowInteraction of Cone
A system of differential equations (9.1.2), (9.2.2), equation of state (9.1.4), energy equation (9.1.5) and the general relations (9.1.6) through (9.1.9) determining enthalpy, entropy, average molecular weight and the speed in a
lissociating and ionizing gas is used for the solution of the problem of the Rowinteraction of a cone taking into account the effect of balanced dissociation and ionization. Here we use the general method of numerical integration of differential equations, as in the case of a gas flow with constant specific heats. The first step in the calculation is the determination of the gas parameters behind the oblique shock for a given angle Bc. The radial component of velocity VCris determined by (9.2.8). The normal component of velocity Vce is determined from the theory of shocks, taking into account the effect of dissociation and ionization (see Chapter 4 Section 2). As a first approximation let us take the value from (4.2.12):
is equal to unity, i.e. we assume the condition of complete stagnation behind the shock so that V,O wO. The pressure corresponding to this condition can be found from formula (4.2.15). Assuming AT,t = 1 and recalling that Mnl = M m sin O,, we get Enthalpy is calculated from the expression (4.2.16) in which it is assumed that ATn = 1 and Vnl= Vmsin 8,:
iL1) = im
7 sin 02, :
It is possible to find density pkl) from the values of p:') and ): using 'i the curves of thermodynamic functions of air [7, 171 and then, as a second approximation, the variation in relative velocity can be determined from (4.2.21)
AF:;) = 1 (pm/pL1)).
From this rise in velocity the improved value of pressure will be pL2)=pm ( 1 + k , M sin28, dFA2)) : and that of enthalpy
(9.3.4) (9.3.5)
Density pi2) can be obtained from the graphs of thermodynamic functions and the value of AT,, is corrected:
4j7A3) 1 (pmIpL2)). =
(9.3.7)
If the value of dFL3) differs by a fairly small quantity from 47;') the
aporoximation procedure can be considered to be complete and the normal, component of velocity behind the shock is determined : vc,= v,, = v,, (1 dFi3))=V sin 8, (1  dFi3)). , (9.3.8) , The tables given in the work [18] can also be used for these calculations. The temperature TL1) and average molecular weight of air (,uav)L1) can be determined from these tables for given values of pi1) and id1) (enthalpy is here denoted by h, m2/sec2). Next, the corresponding value of density can be calculated from formula (4.2.17):
where (pav),=29 gm/gm.mol may be taken for undissociated air. Then pL2), iA2) can be calculated from formulas (9.3.4) through (9.3.6), TL2), (,uav)L2)can be found from the tables and using the expression (9.3.9) the quantity
can be determined. The quantity AFA3)is found from this value of pL2) and formula (9.3.7). Then Vc, = V,, is obtained from the expression (9.3.8). The values of Vr2 and V82 on a neighboring conical surface can be calculated by (9.2.10') and (9.2.1 1') respectively for a given step of integration dB,. Here the derivative (dv~/dB)~ formula (9.2.11') is determined by in (9.2.12), in which the speed of sound a1 can be found from the tables or the graphs of thermodynamic functions for known values of pi2)and id2)(or from p:2) and TL2)). Enthalpy on the selected conical surface is determined by equation (9.1.5): For further calculations it is necessary to use the value of entropy of gas, assuming that the disturbed flow behind the shock is isentropic all over the region and hence that entropy will be everywhere the same as that in the flow immediately behind the shock. This entropy S=Sc=const is determined from the tablesof thermodynamic functionsfor the values of pL2), ii2)(or p p ) , TL2)). The speed of sound az can be found from the tables or graphs for given values of enthalpy i2 and entropy S. The corresponding velocity components
2
V,, and V,, on a conical surface with angle B=8,ZABi can be calculated
is1
by formulas (9.2.10') and (9.2.11'). Similar calculations can be carried out for the surrounding conical surface, etc. The calculations are complete when the
velocity component V,, on one of these surfaces is equal to zero. The corresponding angle (9.2.21) will be the angle of cone under flowinteraction and the corresponding velocity will be the velocity VK (9.2.20) over this cone. The corresponding enthalpy will be
iK = ic ( V :
+  Vg)/2,
ic=iL2).
where V,"=V,:+V;; The temperature TK, pressure p~ and density p~ on the cone can be determined from the tables, graphs or corresponding formulas for enthalpy iKand entropy S=S,. A11 the flow parameters over a cone in a flow of given Mach number M, depend not only on temperature T as characterized in the case of variable , specific heats but also on pressure p, of the undisturbed flow which governs the degree of dissociation and ionization. For given values of parameters PK, M, and T, the flight altitude H may be selected in place of p, as an example of a function determining variation of gas parameters over a conical surface. For given values of T and M, the flow parameters over a cone are functions , of angle pg and flight altitude. Some results of calculations of these parameters are shown in Figs. 9.3.1 through 9.3.3. The nature of the variation in pressure p ~temperature TKand , density p~ over a cone in the presence of dissociation and ionization is the same as that immediately behind the shock. Here the pressure, as behind the shock, varies only slightly during dissociation and ionization. Its value
Fig. 9.3.1. Pressure on cone with adjustment for dissociation 23.5). of gas Tm =220K, MOO=
Fig. 9.3.2. Temperature on cone with adjustment for dissociation of gas (T,=220QK; Mw = 10).
Fig. 9.3.3. Density of air on cone in dissociating flow (Tm =220K, Mm= 10).
basically depends on the condition of undisturbed flow, whereby the maximum change in pressure on a cone cannot exceed some limited value of this pressure (p2PI), which is obtained from (4.2.14) under the condition of Vn2=0, V,I =VI = V, (normal shock) and is equal to pzpl =p,Vk. The temperature and density vary considerably at the same time and this variation is higher with thicker cones. If the cone angles are not high these parameters on the cone are affected only very slightly by physicochemical transformations even at quite high velocities of oncoming flow. For example, calculations show that at M,=24 density hardly changes with altitude up to angle PK=15'. At Mw= 10 (see Fig. 9.3.3) it does not change up to angle PK=35". Thus the interval of angles / 3 ~ corresponds to a comparatively low temperature and a negligibly low that degree of dissociation increases with reduction in speed. A similar phenomenon is observed with reduction in flight altitude. An increase in the thickness of the eC,degree cone brings about more intensive heat 46 , k = 1.4 ing and as a consequence results in '&dissociation and ionization which may 45 considerably affect the flow parameters. The effect of change of shock inclination 8, in front of a cone of angle / 3 =40 is 44 ~ shown in Fig. 9.3.4. The angle 8, diminishes in comparison with that observed at constant specilic heats (k= 1.4). This 43 results in a decrease in the strength of the shock and increase in velocity behind 42 5 1 0 15 20 M , it, which means the velocity over the Fig. 9.3.4. Variation of angle inclination cone increases. ~h~ rise in flow velocity and reduc of shock ahead of cone placed in super(solid linereal tion in sonic velocity lead to an increase sonic flow lineperfect gas).gas; broken in Mach number MK over a conical surface placed in a dissociating gas flow. Gas parameters on a cone in a flow (conditions of flowinteractions: PIC=45", V =6200 m/sec, T, =5800"K, , p,=0.0602 kg/cm2) calculated with and without allowance for dissociation are compared in Table 9.3.1.
._
4. Blunt Cone
4.1 Form of blmt noses The nose section of many flight vehicles has a blunt form. These blunt forms are, first of all, used for very high speeds. Here the main requirement of the nose section is its capacity to withstand the high temperatures of the gas flow over it. However, flight vehicles (or their individual components)
392
AERODYNAMICS
operating at low velocities often have a blunt form which may be required by peculiarities of construction, the type of apparatus, etc. Almost always we have to deal with blunt bodies because it is practically impossible to find absolutely sharp nose sections. This section deals with the conical bodies with different blunt nose profiles that are in most common use. A conical body with a blunt nose of spherical form is shown in Fig. 9.4.1. The equation of the linear part of the generator of this kind of body, tangential to its spherical part, will be
........................................
Conditions of VK, a ~ , TK, calculations mlsec mlsec "K With dissociation Without dissociation
3980 2058 7211
1.934
1.533 x
SO0
(9.4.1)
where rT, XT are the coordinates of a point at the junction of the spherical part and the generator of the cone, which can be determined through the radius of sphere RT and cone angle PKin the form:
The equation of the generator of a spherical nose in the system of coordinates with its origin at the apex will have the nondimensional form:
(9.4.3)
~, where Y ' ~ / RX=X/RT. The angle of inclination /3 of the tangent to the axis at a point onthis part will be given by
tan P=dr/dx=(l
Fig. 9.4.1. Cone with blunt tangential spherical nose.
;)IF
(9.4.4)
The nose section of a cone can be made in such a way that the generator will not be tangential to the spherical surface but will intersect
393
it (Fig. 9.4.2). The radius of this surface drawn for the point with coordinates rT, XT will be R&> RT, where RT is the radius of a tangential nose related to the coordinates r~ and XT by the relations (9.4.2). A cone with a blunt nose in the form of a plane surface (plane section) which can be taken as a sphere of infinite radius is shown in Fig. 9.4.2. The spherical nose and the 6 flat nose can be regarded as particular cases of elliptical surface. A cone with a blunt nose in the form of such a surface is shown in Fig. 9.4.3. The major and minor axes Fig. 9.4.2. Cone with blunt intersecting spherical nose: 2a and 2b of an elliptical nose are its geometrical characterIintersecting spherical nose; 2tangential spherical nose; 3plane section blunt nose. istics where their orientation may be different. A particular case when the major axis is perpendicular to the axis of the cone is shown in Fig. 9.4.3. If the two axes are equal an elliptical nose becomes a spherical one. When one of the axes, coinciding with axis of the cone, is equal to zero the elliptical surface becomes plane. An elliptical nose, like a spherical one, can be tangential or intersect the generator. Besides those considered here, conical bodies with other forms of surface can be used for the nose section. We will not go into the details of these types. The spherical tangential nose and plane section happen to be the most characteristic forms of '0 blunt noses. These may be considered as the limiting cases expres"1 sing other possible forms in their own way. Both of these characteristic forms are relevant to the process of aerodynamic research on blunt bodies with some intermediate form of nose because they allow us to work out the Fig. 9.4.3. Cone with elliptical blunt nose. limiting of the various aerodynamic parameters associated with them. The aerodynamic characteristics of blunt bodies largely depend on the degree of bluntness for a given type of nose. The degree of bluntness
394
AERODYNAMICS
is the ratio of the radius of the base of nose YT to the radius of the maximum cross section of body r,. Moreover the value of the geometrical parameter of the blunt nose itself is very important. For a spherical nose this parameter is the ratio RT/RTC, where RT and RTC are the radii of tangential and intersecting spheres respectively (see Fig. 9.4.2). This parameter varies from unity (for a spherical tangential nose) to zero (for a plane section). For an elliptical nose the ratio of semiaxes b=b/a, which varies in the interval 0 g g .s 1 for a conical blunt body, may be taken as such a parameter. For a given form of spherical or elliptical nose the ratio 2=A/r=, where A is the height of the nose (see Figs. 9.4.2 and 9.4.3), may be chosen as a nondimensional geometrical parameter. If the nose has an elliptical form with the major axis oriented along the normal to the cone axis the value of parameter 2will vary in the interval 0 g AG d. Quantity J=d=0 corresponds to a plane section and the quantity 2=d= (1 sin ~ K ) / c@K corresponds to o~ a tangential spherical nose.
4.2 Special features of supersonic flow  An important aerodynamic property of blunt bodies from the practical point of view is that they heat up less and suffer less damage moving tbrough the atmosphere at very high speeds by comparison with sharpnosed bodies. Let us find out which gas dynamic phenomena are responsible for this property of blunt bodies. A system of flow about a blunt conical body is shown in Fig. 9.4.4. The detached shock of varying intensity at different points of the surface is formed in front of the body. At a distance from the nose the shock wave changes to the usual wave of disturbances of infinitely small strength and an angle of inclination O,=pw=sin1 (l/M,). The maximum strength will be at the apex of the wave (point B in Fig. 9.4.4), where Oc =n/2. As the angle Oc hardly differs from n/2 in the neighborhood of the nose the corresponding section of wave will be quite strong, almost with the strength of a normal shock. The transition of gas particles through such a strong shock will be accompanied by a considerable loss of total head and rise in entropy. As a result the surface of the body is covered by a layer of some thickness in which the gas has high entropy. This high entropy layer is bounded partly by the shock wave and the surface created by rotation of the "sonic" stream line, i.e. the stream line passing through the "sonic" point on the wave (point S having coordinates TST in Fig. 9.4.4). Due to differing degrees of stagnation at different points on the shock wave the flow in the high entropy layer will be characterized by some velocity gradient along normal n as shown in Fig. 9.4.4 and hence by a variable value of local Mach number M across thickness A. If the boundary layer is appreciably thinner than the high entropy layer the velocity gradient in it, by comparison with that in the high entropy layer, can be neglected. This
395
simplifies the analysis. It is assumed that the velocity at different points of the section in the high entropy layer is equal. It can be taken as approximately equal to the mean velocity between the values on the "sonic" and "zero" stream lines passing through the apex of the wave. This velocity is obviously less than the velocity past a sharp cone. The flow region in the neighborhood of the surface of the cone, occupied by the high entropy layer and characterized by low velocities (and hence by a low Mach number and a low Reynolds number), seems to be the deciding factor in the formulation of processes in a boundary layer.
lcbsonic" points; 2shock wave; 36csonic" stream line; 4high entropy layer.
The main peculiarity of the flowinteraction is that the flow regime in the boundary layer changes under the effect of the blunt nose. Due to a reduction in the local Reynolds numbers calculated on the basis of velocity in the high entropy layer the laminar boundary layer changes to a turbulent layer somewhat downstream and so the extent of the laminar boundary layer increases. This helps to reduce friction and heat flows to the wall. Reduction in heat flows based on the enhanced entropy of gas during transition through a shock is known as entropy effect. Here it should be borne in mind that entropy effect not only leads to a reduction in velocity at the external surface of the boundary layer but also to a reduction in the density of the gas, i.e. to a reduction in the Reynolds number. At the same time a rise in entropy also results in an increase (compared to a sharpnosed body) in the temperature at the external limit of the boundary layer. This brings about the antieffect of the high entropy layer whereby there
is some rise in the heat flow from the boundary layer to the wall. However, the resultant entropy effect for an appropriate choice of wall and form of nose leads to a reduction in heat flows, as is shown by calculations and experimental research. The wave drag of a blunt body as a rule increases in comparison to that of a sharpnosed body. However, drag decreases for thin conical bodies with a low degree of bluntness. This is explained by the fact that although pressure rises near the nose it decreases over a considerable part of the surface of the body in comparison with a sharp cone. This phenomenon of reduction in pressure behind the nose is illustrated in Fig. 9.4.5, showing empirical results obtained in aerodynamic wind tunnel tests of thin cones with a blunt nose in the form of a plane section in a supersonic flow of M,=6.85. The minimum pressure i's achieved at a distance of about 10 times the diameter of the nose section. At double this distance from the nose P pressure builds up to its value for a sharp cone. If such a cone is of small length and hence has a small surface having low pressure the reduction in drag for this section will not be enough to compensate for its increase due to the drag increase of the plane section nose. For a lxIDr long enough cone the reduction in 0.0, I 10 100 drag of the peripheral section can be Fig. 9.4.5. Pressure coefficient on surface quite and can in a of cone with plane section blunt nose at ductio ion in the total drag of a bluntMoo=6.85: nosed conical surface compared to by experiment;     calculated by that of a sharpnosed conical surface. conical theory for sharpnosed body. The main point of using a blunt nose is not the reduction in drag, which for a low degree of bluntness seems to be comparatively small, but the appreciable reduction in heat transfer to the wall. This advantage, as shown by experiment, arises mainly in the regime of hypersonic velocities. Other cases of the use of blunt surfaces can be found that are connected with the necessity not so much of reduction in heat transfer as of increase in drag. Space craft have just this form of surface because it is important to have high c, providing most intensive braking in the atmosphere. The great practical relevance of aerodynamic research on blunt bodies is clear from the above discussion.
4.3 Flowinteraction of cone with spherical blunt nose Study of the aerodynamics of a whole blunt body is related to the
397
investigation of the flowinteraction of the front part in the form of a blunt nose of some shbe. The results of these investigations happen to be the basis of calculations of flow parameters over the remaining part of the body. On the other hand these results have their own value as they facilitate determination of the aerodynamic characteristics of a blunt nose. The resultant aerodynamic characteristics of the body can be determined by adding the components for the nose and the remaining part. We must note here that whereas the flowinteraction of a peripheral part of the body depends on its bluntness the flow condition about the nose is determined only by its form or, correctly speaking, by the form of the nose section up to the "sonic" point on its surface. It is obvious that if this ccsonic"point lies on the peripheral surface (downstream behind the line joining nose and body) the disturbances appearing in the region of peripheral surface of the body propagate upstream toward the nose and its flowinteraction should not be examined without consideration of the flow over this region. Let us take the problem of the flowinteraction of the spherical nose of a blunt conical surface. Let us consider this problem as independent, i.e. assume that the peripheral surface of the body has no effect on the flow past the nose. We will study an inviscid flow keeping in mind that the solution sought, irrespective of this limitation, has a great practical significance: it gives the basic conditions of flow outside the boundary layer, the friction and heat transfer governing the processes occurring in the boundary layer itself. First we will study the flow around the point of complete stagnation which is one of the most characteristic points on a spherical surface. Study of this flow is interesting mainly because it is connected with such practical problems as the determination of heat flows, which may reach the highest values. In addition the solution of the problem of flow near the point of complete stagnation helps us to determine the distance from the shock wave to the nose and the distribution of gas dynamic parameters in this small region. Here the solution may be obtained, in general, taking into account the physicochemical transformations of the gas. The equation of motion can be used to solve the problem thus formulated. This equation, according to (3.1.21) for an "inviscid" (V =0) and ccweightless" gas (W= has the following vector form: O), dv/dt =  (l/p) grad p. Let us write equation (9.4.5) in the system of curvilinear orthogonal coordinates. The origin of coordinates of this system coincides with point of complete stagnation 0 on a spherical surface. The x coordinate will be taken along the surface and y coordinate along the normal to it (Fig. 9.4.6). Keeping in mind (3.1.22) and taking a steady flow (aV/at=o), equation (9.4.5) may be given by
grad (V2/2)+rot
F x F= (l/p)gradp.
(9.4.6)
Taking the components of vectors along coordinate lines ql and 92 of the curvilinear system [see (2.4.1 I)] we get equations of motion: $
V x V), =  (l/p) (grad PI,; 1 [grad (V2/2)I2 (rot F x F),=  (llp) (grad p),. +
[grad (V2/21l1+(rot
(9.4.7)
1i
Quantities [grad (V2/2)]1 and [grad (V2/2)]2 can be determined with the help of relation (3.1.24). Replacing p by V2/2, we have
where ql, 92 and q3 are determined by the values of (2.4.40). Since the case of an axisymmetrical flow is being considered we can write a( v2/2)/aq3 a(v2/2)lay=0. = Keeping in view the value of (2.4.42) for Lame's parameters hl and h2, equation (9.4.8) can be written in the form:
( grad$)l
il
+( grad
$)2
i2=(1+&)'
399
From which
v x v=[(rot
From this the components: (rot x Fl1 (rot = V  (rot V)3 V ; 3 2 (rot y x Tl2  (rot P), V (rot 7)3V1. = 3
Since an axisymmetrical flow is being investigated it is necessary to take these expressions. The component of vector (rot p),, by (3.1.28), will be
F, =0 in
or, taking into account (2.4.40) and (2.4.42), it will be given by (rot "= . Therefore (rot V x V), = (rot V), V2
( 1 ;+
)'
{"
ax
[(I+aY
a)
vzl};
]
1
(9.4.10)
7
I I
vy [EL ax
(rot
.[(I+&)
a~
V x Vl2=(rot F),
Vl
From (3.1.24) it follows that 1 aP 1 (grad p)t = . (gradp)2=  .ap hl aqla h2 aq2
(9.4.12)
It is necessary to combine the equation of continuity with the equations of motion. For a steady flow (ap/at=O), according to (2.4.46), this is written in the form
If the region in the neighborhood of the critical point is considered and the oncoming flow has very high supersonic speed the basic equations (9.4.13) through (9.4.15) can be simplified. Actually the flow behind the shock wave under these conditions is incompressible by nature because the Mach number Mz hardly differs from the value [(k 1)/2k]'I2 corresponding to the case of a limiting flow (as M,+ m, k=const) behind a normal shock. Consequently density in the region in question can be assumed constant and equal to p,~, the density behind the shock wave at the point of complete stagnation. Thus the variable density p in the equations of motion and continuity can be replaced by a constant quantity pco. As very high speeds are being considered the shock wave in this case will move nearer to the surface of the body. Here the region of disturbed flow is assumed to be a thin layer of some thickness s which is very small compared to the radius of curvature RT of the surface. T In this way, if s/RT 1 is assumed it is obvious that ~ / R < 1 as 0 <y < s. Hence the quantity ~ / R in comparison of unity can be neglected in equations T of motion (9.4.13), (9.4.14) and equation of continuity (9.4.15). Taking into account the above simplifications and noting that the quantity r in (9.4.15)
can be assumed t s be equal to r E x under the conditions near the points of stagnation, equations (9.4.13) through (9.4.15) will be written in the form:
In eqaations (9.4.16) and (9.4.17) nondimensional density is introduced. Let us show equations (9.4.16) and (9.4.17) can be further simplified. For this purpose we consider the order of magnitude of the terms appearing in these equations. Physically the order of value of Vy will be V, V,o, where Vcois the velocity behind the straight part of a shock wave. The order of magnitude of the coordinate x and y in the immediate vicinity of the stagnation point is determined by the quantities x ~ s oy, N S O (so is the distance from the straight part of a shock wave to the surface of the nose). From equation (9.4.18) it follows that the order of magnitude
so vIw
s Vco o J SO dx,
whence V x VCo.So the order of magnitude of the velocity component V , will be the same as Vy. It can easily be seen that the order of magnitude of the third terms on the left side of the equation will be V:o/ RT and that of the remaining terms V&/s0. Since s0(<RT third terms are of lower order and the can be neglected. In this way equation (9.4.16) and (9.4.17) may be replaced by
The solution of the system of equations (9.4.16'), (9.4.17'), (9.4.18) must satisfy the boundary conditions at any point of the surface of the body at
402
A~RODYNAMICS
y =0 $he normal velocity component V,=O and then at the point of complete ; stagnation, i.e. at y= x =0, the corresponding velocity components V =V =0. , , The boundary conditions immediately behind the shock should also be satisfied by the solution. These conditions, for the velocity components at a point A situated at a distance S from the surface of the nose, will have the form (see Fig. 9.4.6):
(9.4.19) (9.4.20)
where j is the angle of inclination of a tangent surface at a point B on the ? same normal on which point A is situated. The total velocity V, immediately behind the shock is determined with the help of formula (4.3.18). Taking V2= V,, V1=VW and p~/p~wp,/p,o=p in this formula, we have Vc/Vm (cos20, +? sin29c)112. = j (9.4.21) The solution of the problem of flowinteraction in the neighborhood of complete stagnation may lead to the solution of the velocity field. Thus the problem being considered will be a purely kinematic problem. To solve it we have to exclude density and pressure from the equations of motion. For this purpose we differentiate (9.4.16') and (9.4.17') with respect to y and x:
Next we introduce the function determining double the value of the vortex component [component of rotation of velocity (rot V),=2coz; see (2.2.311:
Using the function (9.4.24) and equation of continuity (9.4.18), equations (9.4.22) and (9.4.23) can be transformed. Since the right sides of these equations are identical we get
a v ~ a v ~ ~ ~ + a2vx a2vv ~ ~ a v , +
a da2vx
(axay
ax2 (9.4.25)
aszz
a~ a x a~
X
a2vy aa.
avx +ax
avv3.
a~
Thus the problem lies in finding the solutions of equations (9.4.18) and (9.4.25') satisfying the above boundary conditions. We will find the solution for V , in the neighborhood of a point of complete stagnation with coordinates x=O, y=O in series form: (9.4.26) V,=ao ( y )+a1 ( y ) x + a ~ y ) x2 +a3 (y)x3+ . ., (
in which x is a small parameter and the coefficients a, are some functions of coordinate y. The structure of the series (9.4.26) can be somewhat simplified. In fact the function V , ( x ) will be odd due to symmetry of flow, i.e. the values of velocity components will be equal in magnitude but opposite in sign for the corresponding values of x which are equal in magnitude but opposite in sign. Therefore only the terms with odd powers of x in the series expansion (9.4.26) are maintained, i.e. (9.4.27) VX=al ( y ) x+a3 ( y ) x 3 + . . . . Taking into account that a region in the immediate vicinity of the complete stagnation point is being examined, the terms containing the third or higher power of x can be neglected. In this way, we arrive at the expression Vx=al(y)x. Let us introduce the function
v
(9.4.28)
so that dF/dy =F'(y) =al(y) and F(0) =0. Then Vx=xF'(.v). Substituting the expression for V from (9.4.31) in the equation of continuity , (9.4.18), we get:
From this we find the expression for the other velocity component:
where f (x) is some arbitrary function of x. From the condition of smooth flow over a body Vy (x, 0) =0 and hence, f (x) =  2F(O). But according to (9.4.30) function F(0) =0, so f (x) = 0 and Vy=  2F(y). (9.4.32) To determine the nature of function F(y) we insert (9.4.31) and (9.4.32) in (9.4.25'): [xF"(y)/x] V = VXF1  2F(y)Y1 (y). , (y) According to (9.4.32) the function F(y)gO, so
F1I1 =0. (y)
(9.4.33) (9.4.34)
The general solution of this equation is F(y) =  V,/2 =co cty c2y2.
+ +
. As Vy=0 at y =0 the constant co =0 The other two coefficients can be determined if the conditions on the shock wave near the critical point as x+O are applied. In particular, it follows from the condition (9.4.21) that the velocity immediately behind the "straight" part of the shock wave (0, =4 2 ) at a point C (see Fig. 9.4.6) will be equal to
Vy= vc=pv,. Since the coordinate of this point y=so we get

405
According to this result and equation (9.4.31) the velocity component will be
Vx =X ( C I + 2 ~ 2 ~ ) .
(9.4.36)
Let us take point A on the shock wave at a distance y .s from the surface of the nose section. Equating the velocity V xobtained from (9.4.36)
we get
X(CI+
Taking the limit as x0 and assuming s= so, Vc=>. V,, we get the relation corresponding to the point C on the straight part of the shock wave:
The limit on the left side can be calculated in the following way: It can be seen from Fig. 9.4.6 that the angle (B PC)at a point A on the shock wave is
= Consequently cos (P PC) sin (q +PC). Around the point of complete stagnation the angles p and p, are small and it may be assumed that cos ( p  p c ) q z J + P c . Accordingly
It is obvious that
where
Quantity Rcois the radius of curvature of the shock wave at its apex (if Rc is the variable value of the radius of curvature of the shock wave, Rco=lim R,).
xO +
;
,
(
The angle a, as can be seen from Fig. 9.4.6, is connected with the angle of inclination of shock wave by the formula
o=(7~12) Oc. 
(9.4.39)
To determine (Pc/o)x=owe use formula (4.2.19), which can be rewritten with the help of (4i2.21) and (9.4.39) as poo/pc [tan (6, Pc)]/tan Bc =tan w/tan ( o + PC). = For small values of o and PC pmlpc = o / ( o + PC). Hence
At high speed we can assume that Rco/RTw 1, SO +2 ~ 2 =Voo/RCo. ~0 Along with c,, c2 a third unknown has appeared, the distance so from the shock wave to the nose, in equations (9.4.35) and (9.4.40). Therefore we must add one more independent equation to the system of (9.4.35) and (9.4.40). This additional equation arises from the expression for vortex
407
The relation (9.4:41) can be used to determine the vortex on the surface of the nose as well as immediately behind the shock wave in the flow region around the point of complete stagna